R/CINNA.R
In jafarilab/CINNA: Deciphering Central Informative Nodes in Network Analysis
Defines functions
tsne_centralities
summary_tsne_centralities
summary_calculate_centralities
summary_graph_extract_components
summary_pca_centralities
visualize_dendrogram
visualize_correlations
visualize_heatmap
visualize_pair_correlation
visualize_association
visualize_graph
pca_centralities
calculate_centralities
proper_centralities
giant_component_extract
misc_extract_components
graph_extract_components
Documented in
calculate_centralities
giant_component_extract
graph_extract_components
misc_extract_components
pca_centralities
proper_centralities
summary_calculate_centralities
summary_graph_extract_components
summary_pca_centralities
summary_tsne_centralities
tsne_centralities
visualize_association
visualize_correlations
visualize_dendrogram
visualize_graph
visualize_heatmap
visualize_pair_correlation
#' @title Component extraction of a graph
#'
#' @description This function extracts all connected components of the
#' input, which can be an "igraph" object or a "network" object,
#' and converts them to "igraph" objects.
#' @param x An igraph or a network object.
#' @param directed Whether to create a directed graph (default = TRUE).
#' @param bipartite_proj Whether the bipartite network must be projected or not (default = FALSE).
#' @param num_proj Numbers 1 or 2 that show the number of projects for bipartite graphs (default = 1).
#' @details
#' This function separates different components of an "igraph" or a "network" object and
#' illustrates them as a list of independent graphs. If the input graph was bipartite and the
#' "bipartite_proj" was TRUE, it will project it, and you can decide in which project you want
#' to continue working with.
#' @seealso \code{\link[igraph]{induced.subgraph}}, \code{\link[igraph]{components}}
#' @return A list including the components of the input as igraph objects.
#' The output is a list where each element represents a component and is an igraph object.
#' The components are disconnected subgraphs of the input graph.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @examples
#'
#' data(zachary)
#'
#' graph_extract_components(zachary)
#'
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_bipartite
#' @importFrom igraph bipartite.projection
#' @importFrom igraph simplify
#' @importFrom igraph is_simple
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network network
#' @keywords internal
#' @return A list including the components of the input as igraph objects.
graph_extract_components <- function( x, directed = TRUE, bipartite_proj=FALSE, num_proj=1){
if(!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a
network object")
if(is_igraph(x)){
if( bipartite_proj){
if(is_bipartite(x)){
x<-bipartite.projection(x)[[num_proj]]
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph.splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components<-sapply(1:max(cl$membership), graph.splitting, x = x, cl = cl, simplify = FALSE)
}
}
else{
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components<-sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
}
if( is.network(x)){
edgelist<-as.edgelist(x)
x<-graph_from_edgelist(edgelist, directed = TRUE)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components<-sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
return(components)
}
#' @title Component extraction of miscellaneous graph formats
#'
#' @description This function extracts all components of the input with various formats
#' and converts them to "igraph" objects.
#' @param x The input can be an edgelist, an adjacency matrix, or a graphNEL object.
#' @param directed Whether to create a directed graph. (default = TRUE)
#' @param mode Character scalar, explains how to interpret the supplied matrix.
#' Possible values are: "directed", "undirected", "upper", "lower", "max", "min", "plus". (default = "directed")
#' @param weighted An argument for specifying whether the graph should be weighted or not.
#' If it is NULL, then an unweighted graph is created. (default = NULL)
#' @param unibipartite A boolean parameter describing whether the input edge list corresponds
#' to a bipartite graph. A TRUE value specifies a bipartite graph, and vice versa. (default = FALSE)
#' @param diag Logical scalar, whether to consider the diagonal of the matrix or not.
#' If it is FALSE, then the diagonal is treated as zeros. (default = TRUE)
#' @details
#' This function extracts components from the input object, which can be an edgelist,
#' an adjacency matrix, or a graphNEL object. The result is a list including the components
#' represented as separate graphs.
#' @seealso \code{\link[igraph]{induced.subgraph}}, \code{\link[igraph]{components}},
#' \code{\link[igraph]{graph_from_adjacency_matrix}}
#'
#' @return A list including the components of the input graph as igraph objects.
#' Each element of the list represents a component and is an igraph object.
#' The components are disconnected subgraphs of the input graph.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#' @importFrom igraph graph_from_edgelist
#' @importFrom igraph %>%
#' @importFrom igraph simplify
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom igraph as_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network network
#' @importFrom igraph graph_from_incidence_matrix
#' @importFrom igraph graph_from_adjacency_matrix
#' @importFrom plyr .
misc_extract_components <- function( x ,directed = TRUE, mode = "directed",
weighted = NULL, unibipartite = FALSE,
diag = TRUE){
if (ncol(x)%in%2) {
if (unibipartite%in%FALSE){
x <- graph_from_edgelist(x, directed = directed)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
return(components)
}
else{
el <- cbind(x,1)
incidence_mat <- el[rep(seq_len(nrow(el)), el[,'1']), c(colnames(el)[1], colnames(el)[2])] %>%
{split(.[,colnames(el)[2]], .[,colnames(el)[1]])} %>%
table()
x <- graph_from_incidence_matrix(incidence_mat, directed=directed)
if (!is_simple(x)) x <- simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
return(components)
}
}
if (ncol(x)>2||class(x)%in%"dgCMatrix") {
x <- graph_from_adjacency_matrix(x, mode = mode, weighted = weighted, diag = diag,
add.colnames = NULL, add.rownames = NA)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
return(components)
}
}
#' @title Giant component extraction of a graph
#'
#' @description This function extracts the largest connected component or the giant component
#' of the input graph, which can be an "igraph" object or a "network" object, and converts
#' it to an "igraph" object. For bipartite graphs, this function can perform projection
#' before extracting the components.
#' @param x An igraph or a network object.
#' @param directed Whether to create a directed graph. (default = TRUE)
#' @param bipartite_proj Whether the bipartite network must be projected or not. (default = FALSE)
#' @param num_proj A number indicating the specific projection to work with, especially for
#' bipartite graphs. (default = 1)
#'
#' @details
#' This function identifies the largest component of an "igraph" or a "network" object
#' and returns it as a list containing the giant component as an igraph object and its
#' corresponding edgelist. If the input graph is bipartite and the "bipartite_proj" parameter
#' is set to TRUE, the function can perform projection before extracting the components. You
#' can specify which projection to work with using the "num_proj" parameter.
#' @seealso \code{\link[igraph]{induced.subgraph}}, \code{\link[igraph]{clusters}}
#'
#' @return A list containing the giant component of the input graph. The first element is an igraph object
#' representing the giant component, and the second element is the corresponding edgelist.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Newman, M. (2010). Networks. Oxford University Press.
#' @examples
#' # A graph with 4 vertices
#'
#' data(zachary)
#' giant_component_extract(zachary)
#'
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_bipartite
#' @importFrom igraph bipartite.projection
#' @importFrom igraph is_simple
#' @importFrom igraph simplify
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom igraph as_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network network
giant_component_extract <- function( x, directed = TRUE, bipartite_proj=FALSE ,num_proj=1){
if (!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a network object")
if (is_igraph(x)){
if ( bipartite_proj){
if (is_bipartite(x)){
x <- bipartite.projection(x)[[num_proj]]
if (!is_simple(x)) x <- simplify(x)
cl <- clusters(x)
giant_comp <- induced.subgraph(x, which(cl$membership == which.max(cl$csize)))
giant_comp_edgelist <- as_edgelist(giant_comp, names = TRUE)
}
else stop("The graph is not bipartite")
}
else{
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
giant_comp <- induced.subgraph(x, which(cl$membership == which.max(cl$csize)))
giant_comp_edgelist <- as_edgelist(giant_comp, names = TRUE)
}
}
if ( is.network(x)){
edgelist<-as.edgelist(x)
x <- graph_from_edgelist(edgelist, directed = directed)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
giant_comp <- induced.subgraph(x, which(cl$membership == which.max(cl$csize)))
giant_comp_edgelist <- as_edgelist(giant_comp, names = TRUE)
}
result <- list(giant_comp,giant_comp_edgelist)
return(result)
}
#' @title Proper centrality measure representation
#'
#' @description This function indicates the proper centrality measures of an igraph object
#' based on the network topology.
#' @param x An igraph object.
#' @details
#' This function returns a list including the names of centrality measures that are applicable
#' for the input graph based on its topology.
#' @seealso \code{\link[CINNA]{calculate_centralities}}
#'
#' @return A list including the names of centrality measures that are suitable for the input graph.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @examples
#'
#' data("zachary")
#' proper_centralities(zachary)
#'
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_directed
#' @importFrom igraph is_weighted
proper_centralities<-function(x){
if (!is_igraph(x)) stop(" Error: x must be a class of igraph object ")
if (is_directed(x) && is_weighted(x)){
proper_centralities <- c("Alpha Centrality",
"Burt's Constraint",
"Page Rank" ,
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality",
"Weighted Vertex Degree"
)
}
if (!is_directed(x) && is_weighted(x)) {
proper_centralities<-c( "subgraph centrality scores",
"Topological Coefficient",
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality",
"Weighted Vertex Degree"
)
}
if (!is_directed(x) && !is_weighted(x)) {
proper_centralities <- c( "subgraph centrality scores",
"Topological Coefficient",
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality"
)
}
if (is_directed(x) && !is_weighted(x) ) {
proper_centralities <- c("Alpha Centrality" ,
"Bonacich power centralities of positions" ,
"Page Rank" ,
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality"
)
}
print(proper_centralities)
}
#' @title Centrality measure calculation
#'
#' @description This function computes multitude centrality measures of an igraph object.
#' @param x the component of a network as an igraph object
#' @param except A vector containing names of centrality measures which could be omitted from the calculations.
#' @param include A vector including names of centrality measures which should be computed.
#' @param weights A character scalar specifying the edge attribute to use. (default=NULL)
#' @details
#' This function calculates various types of centrality measures which are applicable to the network topology
#' and returns the results as a list. In the "except" argument, you can specify centrality measures that are
#' not necessary to calculate.
#'
#' @seealso \code{\link[igraph]{alpha.centrality}}, \code{\link[igraph]{bonpow}}, \code{\link[igraph]{constraint}},
#' \code{\link[igraph]{centr_degree}}, \code{\link[igraph]{eccentricity}}, \code{\link[igraph]{eigen_centrality}},
#' \code{\link[igraph]{coreness}}, \code{\link[igraph]{authority_score}}, \code{\link[igraph]{hub_score}},
#' \code{\link[igraph]{transitivity}}, \code{\link[igraph]{page_rank}}, \code{\link[igraph]{betweenness}},
#' \code{\link[igraph]{subgraph.centrality}}, \code{\link[sna]{flowbet}}, \code{\link[sna]{infocent}},
#' \code{\link[sna]{loadcent}}, \code{\link[sna]{stresscent}}, \code{\link[sna]{graphcent}}, \code{\link[centiserve]{topocoefficient}},
#' \code{\link[centiserve]{closeness.currentflow}}, \code{\link[centiserve]{closeness.latora}},
#' \code{\link[centiserve]{communibet}}, \code{\link[centiserve]{communitycent}},
#' \code{\link[centiserve]{crossclique}}, \code{\link[centiserve]{entropy}},
#' \code{\link[centiserve]{epc}}, \code{\link[centiserve]{laplacian}}, \code{\link[centiserve]{leverage}},
#' \code{\link[centiserve]{mnc}}, \code{\link[centiserve]{hubbell}}, \code{\link[centiserve]{semilocal}},
#' \code{\link[centiserve]{closeness.vitality}},
#' \code{\link[centiserve]{closeness.residual}}, \code{\link[centiserve]{lobby}},
#' \code{\link[centiserve]{markovcent}}, \code{\link[centiserve]{radiality}}, \code{\link[centiserve]{lincent}},
#' \code{\link[centiserve]{geokpath}}, \code{\link[centiserve]{katzcent}}, \code{\link[centiserve]{diffusion.degree}},
#' \code{\link[centiserve]{dmnc}}, \code{\link[centiserve]{centroid}}, \code{\link[centiserve]{closeness.freeman}},
#' \code{\link[centiserve]{clusterrank}}, \code{\link[centiserve]{decay}},
#' \code{\link[centiserve]{barycenter}}, \code{\link[centiserve]{bottleneck}}, \code{\link[centiserve]{averagedis}},
#' \code{\link[CINNA]{local_bridging_centrality}}, \code{\link[CINNA]{wiener_index_centrality}}, \code{\link[CINNA]{group_centrality}},
#' \code{\link[CINNA]{dangalchev_closeness_centrality}}, \code{\link[CINNA]{harmonic_centrality}}, \code{\link[igraph]{strength}}
#'
#' @return A list containing centrality measure values. Each column indicates a centrality measure, and each row corresponds to a vertex.
#' The structure of the output is as follows:
#' - The list has named elements, where each element represents a centrality measure.
#' - The value of each element is a numeric vector, where each element of the vector corresponds to a vertex in the network.
#' - The order of vertices in the numeric vectors matches the order of vertices in the input igraph object.
#' - The class of the output is "centrality".
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Bonacich, P., & Lloyd, P. (2001). Eigenvector like measures of centrality for asymmetric relations. Social Networks, 23(3), 191–201.
#'
#' Bonacich, P. (1972). Factoring and weighting approaches to status scores and clique identification. The Journal of Mathematical Sociology, 2(1), 113–120.
#'
#' Bonacich, P. (1987). Power and Centrality: A Family of Measures. American Journal of Sociology, 92(5), 1170–1182.
#'
#' Burt, R. S. (2004). Structural Holes and Good Ideas. American Journal of Sociology, 110(2), 349–399.
#'
#' Batagelj, V., & Zaversnik, M. (2003). An O(m) Algorithm for Cores Decomposition of Networks, 1–9. Retrieved from
#'
#' Seidman, S. B. (1983). Network structure and minimum degree. Social Networks, 5(3), 269–287.
#'
#' Kleinberg, J. M. (1999). Authoritative sources in a hyperlinked environment. Journal of the ACM, 46(5), 604–632.
#'
#' Wasserman, S., & Faust, K. (1994). Social network analysis : methods and applications. American Ethnologist (Vol. 24).
#' Barrat, A., Barthélemy, M., Pastor Satorras, R., & Vespignani, A. (2004). The architecture of complex weighted networks. Proceedings of the National Academy of Sciences of the United States of America , 101(11), 3747–3752.
#'
#' Brin, S., & Page, L. (2010). The Anatomy of a Large Scale Hypertextual Web Search Engine The Anatomy of a Search Engine. Search, 30(June 2000), 1–7.
#'
#' Freeman, L. C. (1978). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239.
#'
#' Brandes, U. (2001). A faster algorithm for betweenness centrality*. The Journal of Mathematical Sociology, 25(2), 163–177.
#'
#' Estrada E., Rodriguez-Velazquez J. A.: Subgraph centrality in Complex Networks. Physical Review E 71, 056103.
#'
#' Freeman, L. C., Borgatti, S. P., & White, D. R. (1991). Centrality in valued graphs: A measure of betweenness based on network flow. Social Networks, 13(2), 141–154.
#'
#' Brandes, U., & Erlebach, T. (Eds.). (2005). Network Analysis (Vol. 3418). Berlin, Heidelberg: Springer Berlin Heidelberg.
#'
#' Stephenson, K., & Zelen, M. (1989). Rethinking centrality: Methods and examples. Social Networks, 11(1), 1–37.
#'
#' Wasserman, S., & Faust, K. (1994). Social network analysis : methods and applications. American Ethnologist (Vol. 24).
#'
#' Brandes, U. (2008). On variants of shortest path betweenness centrality and their generic computation. Social Networks, 30(2), 136–145.
#'
#' Goh, K.-I., Kahng, B., & Kim, D. (2001). Universal Behavior of Load Distribution in Scale Free Networks. Physical Review Letters, 87(27), 278701.
#'
#' Shimbel, A. (1953). Structural parameters of communication networks. The Bulletin of Mathematical Biophysics, 15(4), 501–507.
#'
#' Assenov, Y., Ramrez, F., Schelhorn, S.-E., Lengauer, T., & Albrecht, M. (2008). Computing topological parameters of biological networks. Bioinformatics, 24(2), 282–284.
#'
#' Diekert, V., & Durand, B. (Eds.). (2005). STACS 2005 (Vol. 3404). Berlin, Heidelberg: Springer Berlin Heidelberg.
#'
#' Gräßler, J., Koschützki, D., & Schreiber, F. (2012). CentiLib: comprehensive analysis and exploration of network centralities. Bioinformatics (Oxford, England), 28(8), 1178–9.
#'
#' Latora, V., & Marchiori, M. (2001). Efficient Behavior of Small World Networks. Physical Review Letters, 87(19), 198701.
#'
#' Opsahl, T., Agneessens, F., & Skvoretz, J. (2010). Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32(3), 245–251.
#'
#' Estrada, E., Higham, D. J., & Hatano, N. (2009). Communicability betweenness in complex networks. Physica A: Statistical Mechanics and Its Applications, 388(5), 764–774.
#'
#' Hagberg, Aric, Pieter Swart, and Daniel S Chult. Exploring network structure, dynamics, and function using NetworkX. No. LA-UR-08-05495; LA-UR-08-5495. Los Alamos National Laboratory (LANL), 2008.
#'
#' Kalinka, A. T., & Tomancak, P. (2011). linkcomm: an R package for the generation, visualization, and analysis of link communities in networks of arbitrary size and type. Bioinformatics, 27(14), 2011–2012.
#'
#' Faghani, M. R., & Nguyen, U. T. (2013). A Study of XSS Worm Propagation and Detection Mechanisms in Online Social Networks. IEEE Transactions on Information Forensics and Security, 8(11), 1815–1826.
#'
#' Brandes, U., & Erlebach, T. (Eds.). (2005). Network Analysis (Vol. 3418). Berlin, Heidelberg: Springer Berlin Heidelberg.
#'
#' Lin, C.-Y., Chin, C.-H., Wu, H.-H., Chen, S.-H., Ho, C.-W., & Ko, M.-T. (2008). Hubba: hub objects analyzer--a framework of interactome hubs identification for network biology. Nucleic Acids Research, 36(Web Server), W438–W443.
#'
#' Chin, C., Chen, S., & Wu, H. (2009). cyto Hubba: A Cytoscape Plug in for Hub Object Analysis in Network Biology. Genome Informatics …, 5(Java 5), 2–3.
#'
#' Qi, X., Fuller, E., Wu, Q., Wu, Y., & Zhang, C.-Q. (2012). Laplacian centrality: A new centrality measure for weighted networks. Information Sciences, 194, 240–253.
#'
#' Joyce, K. E., Laurienti, P. J., Burdette, J. H., & Hayasaka, S. (2010). A New Measure of Centrality for Brain Networks. PLoS ONE, 5(8), e12200.
#'
#' Lin, C.-Y., Chin, C.-H., Wu, H.-H., Chen, S.-H., Ho, C.-W., & Ko, M.-T. (2008). Hubba: hub objects analyzer--a framework of interactome hubs identification for network biology. Nucleic Acids Research, 36(Web Server), W438–W443.
#'
#' Hubbell, C. H. (1965). An Input Output Approach to Clique Identification. Sociometry, 28(4), 377.
#'
#' Dangalchev, C. (2006). Residual closeness in networks. Physica A: Statistical Mechanics and Its Applications, 365(2), 556–564.
#'
#' Brandes, U. & Erlebach, T. (2005). Network Analysis: Methodological Foundations, U.S. Government Printing Office.
#'
#' Korn, A., Schubert, A., & Telcs, A. (2009). Lobby index in networks. Physica A: Statistical Mechanics and Its Applications, 388(11), 2221–2226.
#'
#' White, S., & Smyth, P. (2003). Algorithms for estimating relative importance in networks. In Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining KDD ’03 (p. 266). New York, New York, USA: ACM Press.
#'
#' Cornish, A. J., & Markowetz, F. (2014). SANTA: Quantifying the Functional Content of Molecular Networks. PLoS Computational Biology, 10(9), e1003808.
#'
#' Scardoni, G., Petterlini, M., & Laudanna, C. (2009). Analyzing biological network parameters with CentiScaPe. Bioinformatics, 25(21), 2857–2859.
#'
#' Lin, N. (1976). Foundations of Social Research. Mcgraw Hill.
#'
#' Borgatti, S. P., & Everett, M. G. (2006). A Graph theoretic perspective on centrality. Social Networks, 28(4), 466–484.
#'
#' Newman, M. (2010). Networks. Oxford University Press.
#'
#' Junker, Bjorn H., Dirk Koschutzki, and Falk Schreiber(2006). "Exploration of biological network centralities with CentiBiN." BMC bioinformatics 7.1 : 219.
#'
#' Pal, S. K., Kundu, S., & Murthy, C. A. (2014). Centrality measures, upper bound, and influence maximization in large scale directed social networks. Fundamenta Informaticae, 130(3), 317–342.
#'
#' Lin, C.-Y., Chin, C.-H., Wu, H.-H., Chen, S.-H., Ho, C.-W., & Ko, M.-T. (2008). Hubba: hub objects analyzer--a framework of interactome hubs identification for network biology. Nucleic Acids Research, 36(Web Server), W438–W443.
#'
#' Scardoni, G., Petterlini, M., & Laudanna, C. (2009). Analyzing biological network parameters with CentiScaPe. Bioinformatics, 25(21), 2857–2859.
#'
#' Freeman, L. C. (1978). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239.
#'
#' Chen, D.-B., Gao, H., L?, L., & Zhou, T. (2013). Identifying Influential Nodes in Large Scale Directed Networks: The Role of Clustering. PLoS ONE, 8(10), e77455.
#'
#' Jana Hurajova, S. G. and T. M. (2014). Decay Centrality. In 15th Conference of Kosice Mathematicians. Herlany.
#'
#' Viswanath, M. (2009). ONTOLOGY BASED AUTOMATIC TEXT SUMMARIZATION. Vishweshwaraiah Institute of Technology.
#'
#' Przulj, N., Wigle, D. A., & Jurisica, I. (2004). Functional topology in a network of protein interactions. Bioinformatics, 20(3), 340–348.
#'
#' del Rio, G., Koschtzki, D., & Coello, G. (2009). How to identify essential genes from molecular networks BMC Systems Biology, 3(1), 102.
#'
#' Scardoni, G. and Carlo Laudanna, C.B.M.C., 2011. Network centralities for Cytoscape. University of Verona.
#'
#' BOLDI, P. & VIGNA, S. 2014. Axioms for centrality. Internet Mathematics, 00-00.
#'
#' MARCHIORI, M. & LATORA, V. 2000. Harmony in the small-world. Physica A: Statistical Mechanics and its Applications, 285, 539-546.
#'
#' OPSAHL, T., AGNEESSENS, F. & SKVORETZ, J. 2010. Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32, 245-251.
#'
#' OPSAHL, T. 2010. Closeness centrality in networks with disconnected components (http://toreopsahl.com/2010/03/20/closeness-centrality-in-networks-with-disconnected-components/)
#'
#' Michalak, T.P., Aadithya, K.V., Szczepanski, P.L., Ravindran, B. and Jennings, N.R., 2013. Efficient computation of the Shapley value for game-theoretic network centrality. Journal of Artificial Intelligence Research, 46, pp.607-650.
#'
#' Macker, J.P., 2016, November. An improved local bridging centrality model for distributed network analytics. In Military Communications Conference, MILCOM 2016-2016 IEEE (pp. 600-605). IEEE. DOI: 10.1109/MILCOM.2016.7795393
#'
#' DANGALCHEV, C. 2006. Residual closeness in networks. Physica A: Statistical Mechanics and its Applications, 365, 556-564. DOI: 10.1016/j.physa.2005.12.020
#'
#' Alain Barrat, Marc Barthelemy, Romualdo Pastor-Satorras, Alessandro Vespignani: The architecture of complex weighted networks, Proc. Natl. Acad. Sci. USA 101, 3747 (2004)
#'
#' @examples
#' data("zachary")
#' p <- proper_centralities(zachary)
#' calculate_centralities(zachary, include = "Degree Centrality")
#'
#' @export
#'
#' @importFrom igraph is_directed
#' @importFrom igraph is_connected
#' @importFrom igraph is_weighted
#' @importFrom igraph as_edgelist
#' @importFrom stats setNames
#' @importFrom igraph alpha.centrality
#' @importFrom igraph bonpow
#' @importFrom igraph constraint
#' @importFrom igraph centr_degree
#' @importFrom igraph eccentricity
#' @importFrom igraph eigen_centrality
#' @importFrom igraph coreness
#' @importFrom igraph authority_score
#' @importFrom igraph hub_score
#' @importFrom igraph transitivity
#' @importFrom igraph page_rank
#' @importFrom igraph betweenness
#' @importFrom igraph subgraph.centrality
#' @importFrom igraph strength
#' @importFrom network network
#' @importFrom sna flowbet
#' @importFrom sna infocent
#' @importFrom sna loadcent
#' @importFrom sna stresscent
#' @importFrom sna graphcent
#' @importFrom centiserve topocoefficient
#' @importFrom centiserve closeness.currentflow
#' @importFrom centiserve closeness.latora
#' @importFrom centiserve communibet
#' @importFrom centiserve communitycent
#' @importFrom centiserve crossclique
#' @importFrom centiserve entropy
#' @importFrom centiserve epc
#' @importFrom centiserve laplacian
#' @importFrom centiserve leverage
#' @importFrom centiserve mnc
#' @importFrom centiserve hubbell
#' @importFrom centiserve semilocal
#' @importFrom centiserve closeness.vitality
#' @importFrom centiserve closeness.residual
#' @importFrom centiserve lobby
#' @importFrom centiserve markovcent
#' @importFrom centiserve radiality
#' @importFrom centiserve lincent
#' @importFrom centiserve geokpath
#' @importFrom centiserve katzcent
#' @importFrom centiserve diffusion.degree
#' @importFrom centiserve dmnc
#' @importFrom centiserve centroid
#' @importFrom centiserve closeness.freeman
#' @importFrom centiserve clusterrank
#' @importFrom centiserve decay
#' @importFrom centiserve barycenter
#' @importFrom centiserve bottleneck
#' @importFrom centiserve averagedis
calculate_centralities <- function( x, except = NULL, include = NULL, weights = NULL){
if (!("igraph" %in% class(x) && is_connected(x))) stop("The input is not an igraph object
or may not be connected.")
y <- as_edgelist(x)
y <- network(y)
if (is_directed(x) && is_weighted(x) ){
centrality_funcs <- list(
"Alpha Centrality" = function(x) alpha.centrality(x, weights = weights),
"Burt's Constraint" = function(x)constraint(x, weights = weights),
"Page Rank" = function(x)page_rank(x)$vector,
"Average Distance" = function(x)averagedis(x, weights = weights),
"Barycenter Centrality" = function(x)barycenter(x, weights = weights),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = weights),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = weights),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = weights),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = weights)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = weights),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = weights)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = weights)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = weights,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = weights),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = weights),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = weights),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = weights),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = weights),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = weights),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = weights),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = weights),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = weights),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = weights),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = weights),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = weights)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
if (!is_directed(x) && is_weighted(x)){
centrality_funcs <- list(
"subgraph centrality scores" = function(x)subgraph.centrality(x),
"Topological Coefficient" = function(x)topocoefficient(x),
"Average Distance" = function(x)averagedis(x, weights = weights),
"Barycenter Centrality" = function(x)barycenter(x, weights = weights),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = weights),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = weights),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = weights),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = weights)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = weights),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = weights)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = weights)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = weights,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = weights),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = weights),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = weights),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = weights),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = weights),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = weights),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = weights),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = weights),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = weights),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="undirected"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = weights),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = weights),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = weights)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
if(!is_directed(x) && !is_weighted(x)) {
centrality_funcs <- list(
"subgraph centrality scores" = function(x)subgraph.centrality(x),
"Topological Coefficient" = function(x)topocoefficient(x),
"Average Distance" = function(x)averagedis(x, weights = NULL),
"Barycenter Centrality" = function(x)barycenter(x, weights = NULL),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = NULL),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = NULL),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = NULL),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = NULL)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = NULL),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = NULL)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = NULL)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = NULL,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = NULL),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = NULL),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = NULL),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = NULL),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = NULL),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = NULL),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = NULL),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = NULL),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
if(is_directed(x) && !is_weighted(x) ) {
centrality_funcs <- list(
"Alpha Centrality" = function(x) alpha.centrality(x, weights = NULL),
"Bonacich power centralities of positions" = function(x)bonpow(x),
"Page Rank" = function(x)page_rank(x)$vector,
"Average Distance" = function(x)averagedis(x, weights = NULL),
"Barycenter Centrality" = function(x)barycenter(x, weights = NULL),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = NULL),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = NULL),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = NULL),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = NULL)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = NULL),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = NULL)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = NULL)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = NULL,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = NULL),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = NULL),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = NULL),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = NULL),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = NULL),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = NULL),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = NULL),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = NULL),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
}
#' @title PCA Centrality Measures
#'
#' @description This function performs Principal Component Analysis (PCA) on centrality measures.
#' It computes the contributions of variables to the principal components and visualizes them.
#'
#' @param x a list containing the computed centrality values
#' @param scale.unit a boolean constant indicating whether data should be scaled to unit variance (default = TRUE)
#' @param cut.off The intensity that must be exceeded in cumulative percentage of variance of eigen values (default = 80)
#' @param ncp number of dimensions in the final results (default = 5)
#' @param graph a boolean constant indicating whether the graph should be displayed (default = FALSE)
#' @param axes a length 2 vector describing the number of components to plot (default = c(1, 2))
#'
#' @return A plot illustrating the contributions of variables to the principal components.
#' The x-axis represents the centrality measures, and the y-axis represents the contribution.
#' The higher the contribution value, the more important the centrality measure is in the ranking.
#' The plot helps in identifying the most influential centrality measures.
#'
#' @importFrom FactoMineR PCA
#' @importFrom factoextra fviz_contrib
#' @importFrom ggplot2 labs theme_grey theme element_text element_rect
#' @importFrom stats na.omit
#'
#' @examples
#' # Create a data frame with multiple observations
#' centralities <- data.frame(
#' Betweenness = c(0.2, 0.3, 0.5),
#' Closeness = c(0.4, 0.2, 0.6),
#' Degree = c(0.3, 0.1, 0.4),
#' Eigenvector = c(0.1, 0.5, 0.2)
#' )
#' pca_centralities(centralities)
#'
#' @export
pca_centralities <- function(x, scale.unit = TRUE, cut.off = 80, ncp = 2, graph = FALSE, axes = c(1, 2)) {
requireNamespace("FactoMineR", quietly = TRUE)
requireNamespace("factoextra", quietly = TRUE)
requireNamespace("ggplot2", quietly = TRUE)
requireNamespace("utils", quietly = TRUE)
x <- x[!sapply(x, is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
res_pca <- FactoMineR::PCA(x, scale.unit = scale.unit, ncp = ncp, graph = graph, axes = axes)
PCs <- table(res_pca$eig[,"cumulative percentage of variance"] > cut.off)["FALSE"] + 1
if (is.na(PCs)) PCs <- 1
factoextra::fviz_contrib(res_pca, choice = "var", axes = 1:PCs, color = "black", fill = "turquoise") +
ggplot2::labs(x = "\nCentrality measures", y = "Contributions\n") +
ggplot2::theme_grey() +
ggplot2::ggtitle("Contribution of variables via PCA") +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle = 45, vjust = 0.5),
plot.title = ggplot2::element_text(hjust = 0.5),
# Adjust line size here
line = ggplot2::element_line(color = "black", size = 1.5)
)
}
#' @title Graph visualization based on a specific centrality measure
#'
#' @description This function demonstrates the input graph in which the size of nodes
#' indicates calculated centrality value.
#' @param x an igraph object
#' @param computed_centrality_value A vector containing the values of calculated centrality measure for each node.
#' @param centrality.type The type of centrality which should be calculated.
#' @details
#' This function represents the graph in which size of nodes are based on computed centrality value. If the values of wanted centrality
#' measure were computed then by placing them in computed_centrality_value argument to use it for drawing the plot. Otherwise, by only giving the
#' name of favorite centrality measure in centrality.type argument, this function will calculate it and then demonstrates the corresponding graph.
#' @return A plot illustrating the input graph, where the size of nodes represents the calculated centrality values.
#' The function takes an igraph object as input and either a vector of precomputed centrality values or the name of the desired centrality measure.
#' If precomputed centrality values are provided in the computed_centrality_value argument, the graph will be plotted based on those values.
#' If only the name of the desired centrality measure is provided in the centrality.type argument, the function will calculate the centrality values
#' and then plot the graph.
#' The plot shows the nodes of the graph, with their sizes indicating the centrality values. The larger the node, the higher the centrality value.
#' This plot helps visualize the distribution of centrality values across the nodes of the graph.
#' The function returns the plot and does not modify the input graph.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#'
#' @importFrom igraph is.directed
#' @importFrom igraph alpha.centrality
#' @importFrom igraph bonpow
#' @importFrom igraph constraint
#' @importFrom igraph centr_degree
#' @importFrom igraph eccentricity
#' @importFrom igraph eigen_centrality
#' @importFrom igraph coreness
#' @importFrom igraph authority_score
#' @importFrom igraph hub_score
#' @importFrom igraph transitivity
#' @importFrom igraph page_rank
#' @importFrom igraph betweenness
#' @importFrom igraph subgraph.centrality
#' @importFrom igraph strength
#' @importFrom sna flowbet
#' @importFrom sna infocent
#' @importFrom sna loadcent
#' @importFrom sna stresscent
#' @importFrom sna graphcent
#' @importFrom centiserve topocoefficient
#' @importFrom centiserve closeness.currentflow
#' @importFrom centiserve closeness.latora
#' @importFrom centiserve communibet
#' @importFrom centiserve communitycent
#' @importFrom centiserve crossclique
#' @importFrom centiserve entropy
#' @importFrom centiserve epc
#' @importFrom centiserve laplacian
#' @importFrom centiserve leverage
#' @importFrom centiserve mnc
#' @importFrom centiserve hubbell
#' @importFrom centiserve semilocal
#' @importFrom centiserve closeness.vitality
#' @importFrom centiserve closeness.residual
#' @importFrom centiserve lobby
#' @importFrom centiserve markovcent
#' @importFrom centiserve radiality
#' @importFrom centiserve lincent
#' @importFrom centiserve geokpath
#' @importFrom centiserve katzcent
#' @importFrom centiserve diffusion.degree
#' @importFrom centiserve dmnc
#' @importFrom centiserve centroid
#' @importFrom centiserve closeness.freeman
#' @importFrom centiserve clusterrank
#' @importFrom centiserve decay
#' @importFrom centiserve barycenter
#' @importFrom centiserve bottleneck
#' @importFrom centiserve averagedis
#' @importFrom igraph layout_in_circle
#' @importFrom graphics plot
#' @importFrom igraph strength
visualize_graph <- function( x , computed_centrality_value=NULL , centrality.type="Degree Centrality"){
if (is.null(computed_centrality_value)){
if (!("igraph" %in% class(x) && is_connected(x) )) stop("The input is not an igraph object
or may not be connected.")
y <- as_edgelist(x)
y <- network(y)
centrality_funcs <- list(
"subgraph centrality scores"=function(x)subgraph.centrality(x),
"Topological Coefficient"=function(x)topocoefficient(x),
"Alpha Centrality"=function(x) alpha.centrality(x),
"Bonacich power centralities of positions"=function(x)bonpow(x),
"Burt's Constraint"=function(x)constraint(x),
"Page Rank"=function(x)page_rank(x)$vector,
"Average Distance"=function(x)averagedis(x),
"Barycenter Centrality"=function(x)barycenter(x),
"BottleNeck Centrality"=function(x)bottleneck(x),
"Centroid value"=function(x)centroid(x),
"Closeness Centrality (Freeman)"=function(x)closeness.freeman(x),
"ClusterRank"=function(x)clusterrank(x),
"Decay Centrality"=function(x)decay(x),
"Degree Centrality"=function(x)centr_degree(x)$res,
"Diffusion Degree"=function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component"=function(x)dmnc(x),
"Eccentricity Centrality"=function(x)eccentricity(x),
"eigenvector centralities"=function(x)eigen_centrality(x)$vector,
"K-core Decomposition"=function(x)coreness(x),
"Geodesic K-Path Centrality"=function(x)geokpath(x),
"Katz Centrality (Katz Status Index)"=function(x)katzcent(x),
"Kleinberg's authority centrality scores"=function(x)authority_score(x)$vector,
"Kleinberg's hub centrality scores"=function(x)hub_score(x)$vector,
"clustering coefficient"=function(x)transitivity(x,type="local"),
"Lin Centrality"=function(x)lincent(x),
"Lobby Index (Centrality)"=function(x)lobby(x),
"Markov Centrality"=function(x)markovcent(x),
"Radiality Centrality"=function(x)radiality(x),
"Shortest-Paths Betweenness Centrality"=function(x)betweenness(x),
"Current-Flow Closeness Centrality"=function(x)closeness.currentflow(x),
"Closeness centrality (Latora)"=function(x)closeness.latora(x),
"Communicability Betweenness Centrality"=function(x)communibet(x),
"Community Centrality"=function(x)communitycent(x),
"Cross-Clique Connectivity"=function(x)crossclique(x),
"Entropy Centrality"=function(x)entropy(x),
"EPC - Edge Percolated Component"=function(x)epc(x),
"Laplacian Centrality"=function(x)laplacian(x),
"Leverage Centrality"=function(x)leverage(x),
"MNC - Maximum Neighborhood Component"=function(x)mnc(x),
"Hubbell Index"=function(x)hubbell(x),
"Semi Local Centrality"=function(x)semilocal(x),
"Closeness Vitality"=function(x)closeness.vitality(x),
"Residual Closeness Centrality"=function(x)closeness.residual(x),
"Stress Centrality"=function(x)stresscent(y),
"Load Centrality"=function(x)loadcent(y),
"Flow Betweenness Centrality"=function(x)flowbet(y),
"Information Centrality"=function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = NULL)
)
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), centrality.type)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
result <- result[!sapply(result,is.null)]
result <- as.data.frame(result)
result <- scale(result, center = FALSE, scale = TRUE)
plot(x, vertex.size=result*10, layout= layout_in_circle(x,order(result*10) ) ,
vertex.color="turquoise",vertex.frame.color="orange2")
}
else{
computed_centrality_value<-scale(computed_centrality_value, center = FALSE, scale = TRUE)
plot(x, vertex.size=computed_centrality_value*10,
layout= layout_in_circle(x,order(computed_centrality_value*10) ) ,
vertex.color="turquoise",vertex.frame.color="orange2")
}
}
#' @title Pairwise association plot between centrality measures
#'
#' @description This function computes regression between pair of centrality measures
#' to show more details of association among them.
#' @param x a vector containing a centrality measure as independent variable
#' @param y a vector containing a centrality measure as dependent variable
#' @param scale Whether the centrality values should be scaled or not
#' @details
#' This function applies regression analysis on two different centrality values in order to find out the corresponding association
#' between them. Regression analysis is a kind of statitiscal method for approximation the association
#' between variables.It asserts that the value of dependent variable changes when the
#' value of independent variable varies.
#' @return The function returns a regression plot and the results of the regression analysis between two centrality measures.
#' The function takes two vectors, `x` and `y`, which represent the independent and dependent centrality measures, respectively.
#' If the `scale` parameter is set to `TRUE`, the centrality values will be scaled before performing the regression analysis.
#' The function applies regression analysis to estimate the association between the two centrality measures.
#' It creates a scatter plot of the centrality values, with a regression line representing the relationship between the variables.
#' Additionally, the plot includes confidence intervals around the regression line to show the uncertainty of the estimates.
#' The function returns the regression plot and the computed regression results.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Chambers, J. M. (1992). Statistical Models in S. Wadsworth. Pacific Grove, California. Retrieved from
#' [Link to the reference]
#'
#' Wilkinson, G. N., & Rogers, C. E. (1973). Symbolic Description of Factorial Models for Analysis of Variance. Applied Statistics, 22(3), 392.
#'
#' @export
#'
#' @importFrom ggplot2 ggplot
#' @importFrom ggplot2 stat_summary
#' @importFrom ggplot2 geom_smooth
#' @importFrom ggplot2 mean_cl_normal
#' @importFrom ggplot2 aes
#' @importFrom ggplot2 xlab
#' @importFrom ggplot2 ylab
#' @importFrom ggplot2 geom_point
#' @importFrom stats lm
visualize_association <- function( x , y, scale=TRUE){
xname <- substitute(x)
yname <- substitute(y)
df <- as.data.frame(cbind(x,y))
names(df) <- c(xname, yname)
if (scale%in%TRUE){
df <- scale(df, center = TRUE, scale = scale)
df <- as.data.frame(df)
visualization <- ggplot(data=df,aes(x,y)) +
stat_summary(fun.data=mean_cl_normal,geom="point",fill="skyblue3",shape=21,colour="black", size = 3) +
geom_smooth(method='lm',colour = 'red') + xlab(xname) +
ylab(yname)+geom_point(shape=21, fill="blue", color="darkred", size=3)
linear.regression <- lm(df[,2]~df[,1])
return (list(linear.regression=linear.regression, visualization=visualization))
}
else {visualization <- ggplot(data=df,aes(x,y)) +
stat_summary(fun.data=mean_cl_normal,geom="point",fill="skyblue3",shape=21,colour="black", size = 3) +
geom_smooth(method='lm',colour = 'red') + xlab(xname) +
ylab(yname)+geom_point(shape=21, fill="blue", color="darkred", size=3)
linear.regression <- lm(y~x)
return (list(linear.regression=linear.regression, visualization=visualization))}
}
#' @title Pairwise correlation plot between two centrality measures
#'
#' @description This function computes and plots correlation between pair of centrality
#' measures and histogram plots.
#' @param x a vector containing a centrality measure
#' @param y a vector containing another centrality measure
#' @param scale Whether the centrality values should be scaled or not
#' @details
#' This function illustrates the correlation value between two centrality measures and their
#' corresponding scatterplot and histograms.
#' @seealso \code{\link[GGally]{ggpairs}}
#' @return The function computes and plots the correlation between two centrality measures.
#' The function takes two vectors, `x` and `y`, representing the centrality measures to be correlated.
#' If the `scale` parameter is set to `TRUE`, the centrality values will be scaled before computing the correlation.
#' The function creates a correlation plot, which includes a scatterplot of the centrality values and histograms of each measure.
#' The correlation coefficient between the two measures is displayed on the plot.
#' The function returns the correlation plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Emerson, J. W., Green, W. A., Schloerke, B., Crowley, J., Cook, D., Hofmann, H., & Wickham, H. (2013). The Generalized Pairs Plot. Journal of Computational and Graphical Statistics, 22(1), 79–91.
#'
#' @export
#'
#' @importFrom GGally ggpairs
visualize_pair_correlation <- function( x , y, scale=TRUE){
xname <- substitute(x)
yname <- substitute(y)
df <- as.data.frame(cbind(x,y))
names(df) <- c(xname, yname)
if (scale%in%TRUE){
df <- scale(df, center = TRUE, scale = scale)
df <- as.data.frame(df)
ggpairs(df, columns=1:2)
}
else ggpairs(df, columns=1:2)
}
#' @title Heatmap plot between centrality measures
#'
#' @description This function draws heatmap between pair of centrality measures
#' @param x a list indicating calculated centrality measures
#' @param scale Whether the centrality values should be scaled or not
#' @details
#' This function illustrates the heatmap plot of computed centrality measures.
#' @return The function generates a heatmap plot between pairs of centrality measures.
#' The function takes a list x as input, which contains the computed centrality measures.
#' If the scale parameter is set to TRUE, the centrality values will be scaled before plotting the heatmap.
#' The function creates a heatmap plot, where each cell represents the correlation between a pair of centrality measures.
#' The color of each cell indicates the strength and direction of the correlation.
#' The function returns the heatmap plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#'
#' @importFrom pheatmap pheatmap
visualize_heatmap <- function( x, scale = TRUE ){
if(scale%in%TRUE){
x <- x[!sapply(x,is.null)]
d <- as.data.frame(x)
x <- scale(d, center = TRUE, scale = scale)
rownames(x)<- rownames(d)
pheatmap(x, legend = TRUE, cluster_rows = TRUE, cluster_cols = TRUE, fontsize = 8)
}
else{
pheatmap(x, legend = TRUE, cluster_rows = TRUE, cluster_cols = TRUE, fontsize = 8)
}
}
#' @title Correlation plot between centrality measures
#'
#' @description This function draw correlation plot between pair of centrality measures
#' @param x a list indicating calculated centrality measures
#' @param scale Whether the centrality values should be scaled or not(default=TRUE)
#' @param method a character string describing the type of correlation coefficient (or covariance)
#' to be computed. The proper values are "pearson", "kendall", or "spearman". (default="pearson")
#' @details
#' This function illustrates pairwise correlation plot of computed centrality measures.
#' The names of centralities shown in the result plot is abbreviated and compelete names
#' can be seen in "proper_centralities" function.
#' Colors from red to blue indicate the intensity of correlation value. If two centrality
#' measures have an inverse relationship then their correspnding color in plot have to be red
#' and vice versa.
#'
#' @seealso \code{\link[GGally]{ggpairs}}
#' @return The function generates a correlation plot between pairs of centrality measures.
#' The function takes a list x as input, which contains the computed centrality measures.
#' If the scale parameter is set to TRUE, the centrality values will be scaled before computing the correlations.
#' The method parameter specifies the type of correlation coefficient to be computed: "pearson" (default), "kendall", or "spearman".
#' The function creates a pairwise correlation plot, where each cell represents the correlation between a pair of centrality measures.
#' The color of each cell indicates the strength and direction of the correlation, ranging from red (negative correlation) to blue (positive correlation).
#' The function returns the pairwise correlation plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#'
#' @importFrom corrplot corrplot.mixed
#' @importFrom stats cor
visualize_correlations <- function(x, scale=TRUE,method = "pearson"){
if (scale%in%TRUE){
x <- x[!sapply(x,is.null)]
d <- as.data.frame(x)
x <- scale(d, center = TRUE, scale = scale)
name <- unlist(lapply(strsplit(colnames(x), '.', fixed = TRUE), '[', 1))
colnames(x) <- name
M <- cor(x,method = method)
corrplot.mixed(M,tl.cex =0.75)
}
else{
x <- x[!sapply(x,is.null)]
d <- as.data.frame(x)
x <- as.matrix(d)
name <- unlist(lapply(strsplit(colnames(x), '.', fixed = TRUE), '[', 1))
colnames(x) <- name
M <- cor(x,method = method)
corrplot.mixed(M,tl.cex =0.75)
}
}
#' @title Dendrogram plot among centrality measures
#'
#' @description This function demonstrates the vertice dendrogram of a graph based
#' on a centrality type.
#' @param x an igraph object
#' @param centrality.type The type of centrality which should be considered.(default="Degree Centrality")
#' @param computed_centrality_value A vector containing the values of calculated centrality measure for each node.(default=NULL)
#' @param k number of clusters(default=4)
#' @return The function generates a dendrogram plot of a graph's vertices based on a centrality measure.
#' The function takes an igraph object `x` as input, representing the graph.
#' The `centrality.type` parameter specifies the type of centrality measure to be considered (default is "Degree Centrality").
#' If the `computed_centrality_value` parameter is provided, it should be a vector containing the computed centrality values for each node.
#' Otherwise, the function will compute the specified centrality measure for the graph.
#' The `k` parameter specifies the number of clusters to be formed from the dendrogram (default is 4).
#' The function creates a dendrogram plot of the vertices, where the branching structure represents the hierarchical clustering based on the centrality measure.
#' The function returns the dendrogram plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Galili, T. (2015). dendextend: an R package for visualizing, adjusting and comparing trees of hierarchical clustering. Bioinformatics, 31(22), 3718–3720.
#'
#' @export
#'
#' @importFrom igraph alpha.centrality
#' @importFrom igraph bonpow
#' @importFrom igraph constraint
#' @importFrom igraph centr_degree
#' @importFrom igraph eccentricity
#' @importFrom igraph eigen_centrality
#' @importFrom igraph coreness
#' @importFrom igraph authority_score
#' @importFrom igraph hub_score
#' @importFrom igraph transitivity
#' @importFrom igraph page_rank
#' @importFrom igraph betweenness
#' @importFrom igraph subgraph.centrality
#' @importFrom igraph strength
#' @importFrom sna flowbet
#' @importFrom sna infocent
#' @importFrom sna loadcent
#' @importFrom sna stresscent
#' @importFrom sna graphcent
#' @importFrom centiserve topocoefficient
#' @importFrom centiserve closeness.currentflow
#' @importFrom centiserve closeness.latora
#' @importFrom centiserve communibet
#' @importFrom centiserve communitycent
#' @importFrom centiserve crossclique
#' @importFrom centiserve entropy
#' @importFrom centiserve epc
#' @importFrom centiserve laplacian
#' @importFrom centiserve leverage
#' @importFrom centiserve mnc
#' @importFrom centiserve hubbell
#' @importFrom centiserve semilocal
#' @importFrom centiserve closeness.vitality
#' @importFrom centiserve closeness.residual
#' @importFrom centiserve lobby
#' @importFrom centiserve markovcent
#' @importFrom centiserve radiality
#' @importFrom centiserve lincent
#' @importFrom centiserve geokpath
#' @importFrom centiserve katzcent
#' @importFrom centiserve diffusion.degree
#' @importFrom centiserve dmnc
#' @importFrom centiserve centroid
#' @importFrom centiserve closeness.freeman
#' @importFrom centiserve clusterrank
#' @importFrom centiserve decay
#' @importFrom centiserve barycenter
#' @importFrom centiserve bottleneck
#' @importFrom centiserve averagedis
#' @importFrom dendextend circlize_dendrogram
#' @importFrom dendextend set
#' @importFrom dendextend highlight_branches_col
#' @importFrom viridis viridis
#' @importFrom stats dist
#' @importFrom stats hclust
#' @importFrom stats as.dendrogram
visualize_dendrogram <- function( x, centrality.type="Degree Centrality", computed_centrality_value=NULL , k=4){
if (is.null(computed_centrality_value)){
if (!("igraph" %in% class(x) && is_connected(x) )) stop("The input is not an igraph object
or may not be connected.")
y <- as_edgelist(x)
y <- network(y)
centrality_funcs <- list(
"subgraph centrality scores"=function(x)subgraph.centrality(x),
"Topological Coefficient"=function(x)topocoefficient(x),
"Alpha Centrality"=function(x) alpha.centrality(x),
"Bonacich power centralities of positions"=function(x)bonpow(x),
"Burt's Constraint"=function(x)constraint(x),
"Page Rank"=function(x)page_rank(x)$vector,
"Average Distance"=function(x)averagedis(x),
"Barycenter Centrality"=function(x)barycenter(x),
"BottleNeck Centrality"=function(x)bottleneck(x),
"Centroid value"=function(x)centroid(x),
"Closeness Centrality (Freeman)"=function(x)closeness.freeman(x),
"ClusterRank"=function(x)clusterrank(x),
"Decay Centrality"=function(x)decay(x),
"Degree Centrality"=function(x)centr_degree(x)$res,
"Diffusion Degree"=function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component"=function(x)dmnc(x),
"Eccentricity Centrality"=function(x)eccentricity(x),
"eigenvector centralities"=function(x)eigen_centrality(x)$vector,
"K-core Decomposition"=function(x)coreness(x),
"Geodesic K-Path Centrality"=function(x)geokpath(x),
"Katz Centrality (Katz Status Index)"=function(x)katzcent(x),
"Kleinberg's authority centrality scores"=function(x)authority_score(x)$vector,
"Kleinberg's hub centrality scores"=function(x)hub_score(x)$vector,
"clustering coefficient"=function(x)transitivity(x,type="local"),
"Lin Centrality"=function(x)lincent(x),
"Lobby Index (Centrality)"=function(x)lobby(x),
"Markov Centrality"=function(x)markovcent(x),
"Radiality Centrality"=function(x)radiality(x),
"Shortest-Paths Betweenness Centrality"=function(x)betweenness(x),
"Current-Flow Closeness Centrality"=function(x)closeness.currentflow(x),
"Closeness centrality (Latora)"=function(x)closeness.latora(x),
"Communicability Betweenness Centrality"=function(x)communibet(x),
"Community Centrality"=function(x)communitycent(x),
"Cross-Clique Connectivity"=function(x)crossclique(x),
"Entropy Centrality"=function(x)entropy(x),
"EPC - Edge Percolated Component"=function(x)epc(x),
"Laplacian Centrality"=function(x)laplacian(x),
"Leverage Centrality"=function(x)leverage(x),
"MNC - Maximum Neighborhood Component"=function(x)mnc(x),
"Hubbell Index"=function(x)hubbell(x),
"Semi Local Centrality"=function(x)semilocal(x),
"Closeness Vitality"=function(x)closeness.vitality(x),
"Residual Closeness Centrality"=function(x)closeness.residual(x),
"Stress Centrality"=function(x)stresscent(y),
"Load Centrality"=function(x)loadcent(y),
"Flow Betweenness Centrality"=function(x)flowbet(y),
"Information Centrality"=function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = NULL)
)
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), centrality.type)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
result <-result[!sapply(result,is.null)]
result <- as.data.frame(result)
rownames(result) = V(x)$name
result <- scale(result, center = TRUE, scale = TRUE)
dend <- result%>% dist %>% hclust %>% as.dendrogram %>%
highlight_branches_col(viridis(100)) %>%
set("branches_k_color", k=3)%>%
set("labels_colors")%>%set("nodes_pch", 20)
circlize_dendrogram(dend, labels_track_height = NA, dend_track_height = .4)
}
else{
computed_centrality_value <- scale(computed_centrality_value, center = FALSE, scale = TRUE)
names(computed_centrality_value) = V(x)$name
dend <- computed_centrality_value%>% dist %>% hclust %>% as.dendrogram %>%
highlight_branches_col(viridis(100)) %>% set("branches_k_color", k=3)%>%
set("labels_colors")%>%set("nodes_pch", 20)
circlize_dendrogram(dend, labels_track_height = NA, dend_track_height = .4)
}
}
#' @title Summarize PCA result related to centrality measures
#'
#' @description This function summarizes the PCA result related to centrality measures.
#' @param x A list containing the computed centrality values.
#' @param scale.unit A boolean value indicating whether the data should be scaled to unit variance (default = TRUE).
#' @param ncp The number of dimensions in the final results (default = 5).
#' @return The result of the \code{pca_centralities} function, which includes the PCA analysis results such as eigenvalues, variance explained, and scores.
#' The returned value is an object of class "PCA" from the FactoMineR package. It contains the following components:
#' \item{eig}{A numeric vector of eigenvalues, indicating the amount of variance explained by each principal component.}
#' \item{var}{A numeric vector of proportions of variance explained by each principal component.}
#' \item{ind}{A data frame of individual scores, where each row represents an individual and each column represents a principal component.}
#' \item{call}{The function call used to create the PCA object.}
#' \item{call2}{The call used to compute the PCA analysis.}
#' \item{call3}{The call used to project the individuals.}
#' \item{svd}{The singular value decomposition of the data matrix.}
#' \item{cor}{The correlation matrix of the variables.}
#' \item{cos2}{The squared cosines of the variables.}
#' \item{contrib}{The contributions of the variables to the principal components.}
#' \item{proj}{The projected coordinates of the individuals on the principal components.}
#' \item{quali.sup}{The results of the supplementary qualitative variables analysis.}
#' \item{quali.sup.ind}{The results of the supplementary individuals analysis.}
#' \item{quanti.sup}{The results of the supplementary quantitative variables analysis.}
#' \item{quanti.sup.var}{The results of the supplementary variables analysis.}
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom FactoMineR PCA
#' @importFrom stats na.omit
summary_pca_centralities <- function( x , scale.unit = TRUE,ncp = 5){
x <- x[!sapply(x,is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
res_pca <- PCA(x, scale.unit = scale.unit, ncp = ncp, graph = FALSE)
l <- list(eigen= res_pca$eig, contribution= res_pca$var$contrib)
print(l)
}
#' @title Summarize component extraction of a graph
#'
#' @description This function summarizes all components of the input which can be an "igraph" object or a "network" object.
#' @param x An igraph or a network object.
#' @param directed A boolean value indicating whether to create a directed graph (default = TRUE).
#' @param bipartite_proj A boolean value indicating whether the bipartite network should be projected (default = FALSE).
#' @param num_proj A number indicating the number of projects specifically for bipartite graphs (default = 1).
#' @return A list of igraph objects representing the extracted components of the graph.
#' Each element in the list corresponds to a component, and each component is represented by an igraph object.
#' The components are extracted based on the connectivity of the graph and can include single nodes, disconnected subgraphs, or connected subgraphs.
#' If the graph is directed and the \code{directed} parameter is set to FALSE, the function returns the weakly connected components.
#' If the graph is bipartite and the \code{bipartite_proj} parameter is set to TRUE, the function returns the projected graph components.
#' If the graph is neither directed nor bipartite, the function returns the connected components.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_bipartite
#' @importFrom igraph bipartite.projection
#' @importFrom igraph is_simple
#' @importFrom igraph simplify
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network is.network
#' @importFrom network network
#' @importFrom stats na.omit
summary_graph_extract_components <- function( x, directed = TRUE, bipartite_proj = FALSE , num_proj = 1){
if (!("igraph" %in% class(x) || ("network" %in% class(x) ))) stop("The input is not an igraph or a network object")
if (is_igraph(x)) {
if (bipartite_proj){
if (is_bipartite(x)){
x <- bipartite.projection(x)[[num_proj]]
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
}
else{
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
}
if( is.network(x)){
edgelist <- as.edgelist(x)
x <- graph_from_edgelist(edgelist, directed = directed)
if (!is_simple(x)) gr <- simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
summary.list <- lapply(components, function(x) summary(x))
}
#' @title Summarize centrality measure calculation results
#' @description This function computes the minimum, first quartile, median,
#' mean, third quartile, and maximum values of the computed centrality measures.
#' @param x Centrality measure calculation results, typically a numeric vector.
#' @return A list summarizing the computed centrality measures. The list includes the following elements:
#' - "minimum": The minimum value of the centrality measures.
#' - "first_quartile": The first quartile (25th percentile) value of the centrality measures.
#' - "median": The median (50th percentile) value of the centrality measures.
#' - "mean": The mean value of the centrality measures.
#' - "third_quartile": The third quartile (75th percentile) value of the centrality measures.
#' - "maximum": The maximum value of the centrality measures.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom stats na.omit
summary_calculate_centralities <- function(x){
x <- x[!sapply(x, is.null)]
x <- na.omit(x)
summary_list <- lapply(x, function(x) summary(x))
return(summary_list)
}
#' @title Summarize t-Distributed Stochastic Neighbor Embedding (t-SNE) on centrality measures
#'
#' @description This function summarizes the t-SNE analysis results on centrality measures.
#' @param x A list containing the computed centrality values.
#' @param dims An integer specifying the number of output dimensions (default = 2).
#' @param perplexity A numeric value representing a flexible measure of the efficient number of neighbors.
#' The performance of t-SNE is fairly robust to changes in perplexity, and typical values are between 5 and 50 (default = 5).
#' @param scale A logical value indicating whether the centrality values should be scaled or not (default = TRUE).
#' @seealso \code{\link[Rtsne]{Rtsne}}
#' @return A list containing the following elements:
#' - "Y": A matrix containing the new representations for the objects after t-SNE.
#' - "costs": The cost for every object after the final iteration of t-SNE.
#' The "costs" provide information about the optimization process.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom Rtsne Rtsne
#' @importFrom stats na.omit
summary_tsne_centralities<-function( x , dims = 2, perplexity = 5, scale = TRUE){
x <- x[!sapply(x, is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
if (scale%in%TRUE){
x <- scale(x, center = TRUE, scale = scale)
x <- x[!duplicated(x), ]
tsne.Y <- Rtsne(t(x), dims = dims, perplexity = perplexity, check_duplicates = FALSE)$Y
rownames(tsne.Y)<-colnames(x)
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity ,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
tsne_cost <- sort(cost)
res_tsne <- list(tsne.Y = tsne.Y, tsne_cost = tsne_cost)
return(res_tsne)
}
else{
x <- x[!duplicated(x), ]
tsne.Y <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$Y
rownames(tsne.Y) <- colnames(x)
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
tsne_cost <- sort(cost)
res_tsne <- list(tsne.Y = tsne.Y,tsne_cost = tsne_cost)
return(res_tsne)
}
}
#' @title t-Distributed Stochastic Neighbor Embedding (t-SNE) on centrality measures
#'
#' @description This function applies t-SNE, a dimensionality reduction algorithm, to centrality measures.
#' @param x A list containing the computed centrality values.
#' @param dims An integer specifying the number of output dimensions (default = 2).
#' @param perplexity A numeric value representing a flexible measure of the efficient number of neighbors.
#' The performance of t-SNE is fairly robust to changes in perplexity, and typical values are between 5 and 50 (default = 5).
#' @param scale A logical value indicating whether the centrality values should be scaled or not (default = TRUE).
#' @details t-SNE is a non-linear dimensionality reduction algorithm used for exploring high-dimensional data. It maps multi-dimensional centrality measure data to a lower-dimensional space suitable for analysis and visualization.
#' @seealso \code{\link[Rtsne]{Rtsne}}
#' @return A cost plot of t-SNE results, which displays centralities in order of their corresponding costs.
#' The cost plot provides information about the optimization process and the quality of the embedding.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' van der Maaten, L. (2014). Accelerating t-SNE using Tree-Based Algorithms. Journal of Machine Learning Research, 15, 3221–3245.
#' Van Der Maaten, L. J. P., & Hinton, G. E. (2008). Visualizing High-Dimensional Data Using t-SNE. Journal of Machine Learning Research, 9, 2579–2605.
#' @export
#' @importFrom Rtsne Rtsne
#' @importFrom ggplot2 geom_bar
#' @importFrom ggplot2 guides
#' @importFrom ggplot2 xlab
#' @importFrom ggplot2 ylab
#' @importFrom ggplot2 ggtitle
#' @importFrom ggplot2 theme
#' @importFrom stats na.omit
#' @importFrom stats reorder
#' @return A cost plot of t-SNE results, which displays centralities in order of their corresponding costs.
#' The cost plot is a ggplot object that represents the optimization process and the quality of the embedding.
#' The x-axis represents the iterations of the t-SNE algorithm, and the y-axis represents the cost associated with each iteration.
#' The cost measures the discrepancy between the original high-dimensional space and the low-dimensional embedding.
#' By examining the cost plot, you can assess the convergence and stability of the t-SNE algorithm and evaluate the quality of the embedding.
tsne_centralities <- function( x , dims = 2, perplexity = 5, scale = TRUE){
x <- x[!sapply(x,is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
if (scale%in%TRUE){
x <- scale(x, center = TRUE, scale = scale)
x <- x[!duplicated(x), ]
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
df <- data.frame(sort(cost))
ggplot(data=df, aes(x=reorder(rownames(df), -cost), y=cost, fill=rownames(df))) +
geom_bar(colour="black", fill="turquoise", width=.8, stat="identity") +
guides(fill=FALSE) +
xlab("Centrality measures") + ylab("tsne costs") +
ggtitle("tsne cost results")+ theme(axis.text.x=element_text(angle=45, vjust=0.5),plot.title = element_text(hjust = 0.5)
,panel.border = element_rect(colour = "black", fill=NA, size=1.5))
}
else{
x <- x[!duplicated(x), ]
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
df <- data.frame(sort(cost))
ggplot(data=df, aes(x=reorder(rownames(df), -cost), y=cost, fill=rownames(df))) +
geom_bar(colour="black", fill="turquoise", width=.8, stat="identity") +
guides(fill=FALSE) +
xlab("Centrality measures") + ylab("tsne costs") +
ggtitle("tsne cost results")+ theme(axis.text.x=element_text(angle=45, vjust=0.5),
plot.title = element_text(hjust = 0.5)
,panel.border = element_rect(colour = "black", fill=NA, size=1.5))
}
}
#' @title Dangalchev Closeness Centrality
#'
#' @description This function computes the Dangalchev Closeness Centrality for nodes in a network.
#' The Dangalchev Closeness Centrality measures closeness by removing nodes and edges, allowing for easier evaluation and handling of unconnected graphs.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#' @param mode A character value indicating whether the shortest paths "in" or "out" of the nodes in directed graphs should be considered. For undirected graphs, use "all".
#' @param weights A numeric vector indicating the weights of the edges.
#'
#' @seealso \code{\link[centiserve]{closeness.residual}}
#' @return A numeric vector including the centrality values for each node.
#' The centrality values represent the Dangalchev Closeness Centrality measure for each node in the network.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Dangalchev, C. (2006). Residual closeness in networks. Physica A: Statistical Mechanics and its Applications, 365, 556-564. DOI: 10.1016/j.physa.2005.12.020
#'
#' @examples
#'
#' data(zachary)
#'
#' dangalchev_closeness_centrality(zachary)
#'
#' @export
#' @importFrom igraph distances
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
#' @importFrom intergraph asIgraph
dangalchev_closeness_centrality<-function (x, vids = V(x), mode = c("all", "out", "in"), weights = NULL){
if (!("igraph" %in% class(x) || "network" %in% class(x))) stop("The input is not an igraph or a
network object")
if (is_igraph(x)){
distMat<-1/2^distances(x)
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
distMat<-1/2^distances(x)
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
return(res)
}
#' @title Local Bridging Centrality
#'
#' @description This function computes the Local Bridging Centrality for nodes in a network. The Local Bridging Centrality classifies nodes based on their structural links among the dense components.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#'
#' @seealso \code{\link[igraph]{betweenness}}
#' @return A numeric vector including the centrality values for each node.
#' The centrality values represent the Local Bridging Centrality measure for each node in the network.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Macker, J.P. (2016). An improved local bridging centrality model for distributed network analytics. In Military Communications Conference, MILCOM 2016-2016 IEEE (pp. 600-605). IEEE. DOI: 10.1109/MILCOM.2016.7795393
#'
#' @examples
#'
#' data(zachary)
#'
#' local_bridging_centrality(zachary)
#'
#' @export
#' @importFrom igraph degree
#' @importFrom igraph neighbors
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
#' @importFrom intergraph asIgraph
local_bridging_centrality<-function (x, vids = V(x)){
if (!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a
network object")
f <- function(v){
(1/degree(x,v))/sum(1/degree(x,neighbors(x,v,'all')))}
if (is_igraph(x)){
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
return(results)
}
#' @title Wiener Index Centrality
#'
#' @description This function computes the Wiener Index Centrality for nodes in a network. The Wiener index is a measure of centrality that computes the sum of all shortest paths between a node and all other related nodes in the graph. In essence, it is similar to closeness centrality, but without taking the reciprocal, resulting in a measure with the opposite meaning.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#' @param mode A character value indicating whether the shortest paths "in" or "out" of the nodes in directed graphs should be considered. For undirected graphs, "all" is used.
#' @param weights Numeric vector indicating weights of the edges.
#'
#' @return A numeric vector including the centrality values for each node.
#' The centrality values represent the Wiener Index Centrality measure for each node in the network.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Scardoni, G. and Carlo Laudanna, C.B.M.C. (2011). Network centralities for Cytoscape. University of Verona.
#'
#' @examples
#'
#' data(zachary)
#'
#' wiener_index_centrality(zachary)
#'
#' @export
#' @importFrom igraph shortest.paths
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
#' @importFrom intergraph asIgraph
wiener_index_centrality<-function (x, vids = V(x), mode = c("all", "out", "in"), weights = NULL){
if (!("igraph" %in% class(x)||"network" %in% class(x))) stop("The input is not an igraph or a
network object")
if (is_igraph(x)){
distMat<- shortest.paths(x , mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
distMat<- shortest.paths(x , mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
return(res=res)
}
#' @title Harmonic Centrality
#'
#' @description This function computes the Harmonic Centrality for nodes in a network. The harmonic centrality metric is defined as the denormalized reciprocal of the harmonic mean of all distances.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#' @param mode A character value, indicating the type of degree to consider ("out" for out-degree, "in" for in-degree, "total" for the sum of the two). For undirected graphs, this argument is ignored. The default value is "total".
#' @param weights Numeric vector indicating weights of the edges.
#'
#' @return A numeric vector of centrality values for each node. The length of the vector is equal to the number of nodes in the network.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' BOLDI, P. & VIGNA, S. 2014. Axioms for centrality. Internet Mathematics, 00-00.
#'
#' MARCHIORI, M. & LATORA, V. 2000. Harmony in the small-world. Physica A: Statistical Mechanics and its Applications, 285, 539-546.
#'
#' OPSAHL, T., AGNEESSENS, F. & SKVORETZ, J. 2010. Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32, 245-251.
#'
#' OPSAHL, T. 2010. Closeness centrality in networks with disconnected components (http://toreopsahl.com/2010/03/20/closeness-centrality-in-networks-with-disconnected-components/)
#'
#' @examples
#'
#' data(zachary)
#'
#' harmonic_centrality(zachary)
#'
#' @export
#' @importFrom igraph distances
#' @importFrom igraph vcount
#' @importFrom intergraph asIgraph
#' @importFrom igraph degree
#' @importFrom igraph neighbors
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
harmonic_centrality <- function (x, vids = V(x), mode = c("all", "out", "in"), weights = NULL){
if (!("igraph" %in% class(x) || "network" %in% class(x))) stop("The input is not an igraph or a
network object")
if (is_igraph(x)){
distMat<-1/distances(x, mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
distMat<-1/distances(x, mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
return(res)
}
#' @title Group Centrality
#'
#' @description This function computes the Group Centrality for nodes in a network. Group centrality considers a consistent ranking of each node to be calculated, taking into account the diverse possible synergies among possible groups of vertices.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#'
#' @return A numeric vector of centrality values for each node. The length of the vector is equal to the number of nodes in the network.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Michalak, T.P., Aadithya, K.V., Szczepanski, P.L., Ravindran, B. and Jennings, N.R., 2013. Efficient computation of the Shapley value for game-theoretic network centrality. Journal of Artificial Intelligence Research, 46, pp.607-650.
#'
#' https://www.civilica.com/Paper-IBIS07-IBIS07_127.html
#'
#' @examples
#'
#' data(zachary)
#'
#' group_centrality(zachary)
#'
#' @export
#' @importFrom igraph neighbors
#' @importFrom igraph degree
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
group_centrality<-function (x, vids = V(x)){
if (!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a
network object")
f <- function(v){ 1/(1 + degree(x,v)) + sum(1/(1+degree(x,neighbors(x,v,'all'))))}
if (is_igraph(x)){
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
return(results)
}
jafarilab/CINNA documentation built on Aug. 19, 2023, 4:49 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions tsne_centralities summary_tsne_centralities summary_calculate_centralities summary_graph_extract_components summary_pca_centralities visualize_dendrogram visualize_correlations visualize_heatmap visualize_pair_correlation visualize_association visualize_graph pca_centralities calculate_centralities proper_centralities giant_component_extract misc_extract_components graph_extract_components
Documented in calculate_centralities giant_component_extract graph_extract_components misc_extract_components pca_centralities proper_centralities summary_calculate_centralities summary_graph_extract_components summary_pca_centralities summary_tsne_centralities tsne_centralities visualize_association visualize_correlations visualize_dendrogram visualize_graph visualize_heatmap visualize_pair_correlation
#' @title Component extraction of a graph
#'
#' @description This function extracts all connected components of the
#' input, which can be an "igraph" object or a "network" object,
#' and converts them to "igraph" objects.
#' @param x An igraph or a network object.
#' @param directed Whether to create a directed graph (default = TRUE).
#' @param bipartite_proj Whether the bipartite network must be projected or not (default = FALSE).
#' @param num_proj Numbers 1 or 2 that show the number of projects for bipartite graphs (default = 1).
#' @details
#' This function separates different components of an "igraph" or a "network" object and
#' illustrates them as a list of independent graphs. If the input graph was bipartite and the
#' "bipartite_proj" was TRUE, it will project it, and you can decide in which project you want
#' to continue working with.
#' @seealso \code{\link[igraph]{induced.subgraph}}, \code{\link[igraph]{components}}
#' @return A list including the components of the input as igraph objects.
#' The output is a list where each element represents a component and is an igraph object.
#' The components are disconnected subgraphs of the input graph.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @examples
#'
#' data(zachary)
#'
#' graph_extract_components(zachary)
#'
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_bipartite
#' @importFrom igraph bipartite.projection
#' @importFrom igraph simplify
#' @importFrom igraph is_simple
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network network
#' @keywords internal
#' @return A list including the components of the input as igraph objects.
graph_extract_components <- function( x, directed = TRUE, bipartite_proj=FALSE, num_proj=1){
if(!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a
network object")
if(is_igraph(x)){
if( bipartite_proj){
if(is_bipartite(x)){
x<-bipartite.projection(x)[[num_proj]]
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph.splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components<-sapply(1:max(cl$membership), graph.splitting, x = x, cl = cl, simplify = FALSE)
}
}
else{
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components<-sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
}
if( is.network(x)){
edgelist<-as.edgelist(x)
x<-graph_from_edgelist(edgelist, directed = TRUE)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components<-sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
return(components)
}
#' @title Component extraction of miscellaneous graph formats
#'
#' @description This function extracts all components of the input with various formats
#' and converts them to "igraph" objects.
#' @param x The input can be an edgelist, an adjacency matrix, or a graphNEL object.
#' @param directed Whether to create a directed graph. (default = TRUE)
#' @param mode Character scalar, explains how to interpret the supplied matrix.
#' Possible values are: "directed", "undirected", "upper", "lower", "max", "min", "plus". (default = "directed")
#' @param weighted An argument for specifying whether the graph should be weighted or not.
#' If it is NULL, then an unweighted graph is created. (default = NULL)
#' @param unibipartite A boolean parameter describing whether the input edge list corresponds
#' to a bipartite graph. A TRUE value specifies a bipartite graph, and vice versa. (default = FALSE)
#' @param diag Logical scalar, whether to consider the diagonal of the matrix or not.
#' If it is FALSE, then the diagonal is treated as zeros. (default = TRUE)
#' @details
#' This function extracts components from the input object, which can be an edgelist,
#' an adjacency matrix, or a graphNEL object. The result is a list including the components
#' represented as separate graphs.
#' @seealso \code{\link[igraph]{induced.subgraph}}, \code{\link[igraph]{components}},
#' \code{\link[igraph]{graph_from_adjacency_matrix}}
#'
#' @return A list including the components of the input graph as igraph objects.
#' Each element of the list represents a component and is an igraph object.
#' The components are disconnected subgraphs of the input graph.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#' @importFrom igraph graph_from_edgelist
#' @importFrom igraph %>%
#' @importFrom igraph simplify
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom igraph as_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network network
#' @importFrom igraph graph_from_incidence_matrix
#' @importFrom igraph graph_from_adjacency_matrix
#' @importFrom plyr .
misc_extract_components <- function( x ,directed = TRUE, mode = "directed",
weighted = NULL, unibipartite = FALSE,
diag = TRUE){
if (ncol(x)%in%2) {
if (unibipartite%in%FALSE){
x <- graph_from_edgelist(x, directed = directed)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
return(components)
}
else{
el <- cbind(x,1)
incidence_mat <- el[rep(seq_len(nrow(el)), el[,'1']), c(colnames(el)[1], colnames(el)[2])] %>%
{split(.[,colnames(el)[2]], .[,colnames(el)[1]])} %>%
table()
x <- graph_from_incidence_matrix(incidence_mat, directed=directed)
if (!is_simple(x)) x <- simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
return(components)
}
}
if (ncol(x)>2||class(x)%in%"dgCMatrix") {
x <- graph_from_adjacency_matrix(x, mode = mode, weighted = weighted, diag = diag,
add.colnames = NULL, add.rownames = NA)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
return(components)
}
}
#' @title Giant component extraction of a graph
#'
#' @description This function extracts the largest connected component or the giant component
#' of the input graph, which can be an "igraph" object or a "network" object, and converts
#' it to an "igraph" object. For bipartite graphs, this function can perform projection
#' before extracting the components.
#' @param x An igraph or a network object.
#' @param directed Whether to create a directed graph. (default = TRUE)
#' @param bipartite_proj Whether the bipartite network must be projected or not. (default = FALSE)
#' @param num_proj A number indicating the specific projection to work with, especially for
#' bipartite graphs. (default = 1)
#'
#' @details
#' This function identifies the largest component of an "igraph" or a "network" object
#' and returns it as a list containing the giant component as an igraph object and its
#' corresponding edgelist. If the input graph is bipartite and the "bipartite_proj" parameter
#' is set to TRUE, the function can perform projection before extracting the components. You
#' can specify which projection to work with using the "num_proj" parameter.
#' @seealso \code{\link[igraph]{induced.subgraph}}, \code{\link[igraph]{clusters}}
#'
#' @return A list containing the giant component of the input graph. The first element is an igraph object
#' representing the giant component, and the second element is the corresponding edgelist.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Newman, M. (2010). Networks. Oxford University Press.
#' @examples
#' # A graph with 4 vertices
#'
#' data(zachary)
#' giant_component_extract(zachary)
#'
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_bipartite
#' @importFrom igraph bipartite.projection
#' @importFrom igraph is_simple
#' @importFrom igraph simplify
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom igraph as_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network network
giant_component_extract <- function( x, directed = TRUE, bipartite_proj=FALSE ,num_proj=1){
if (!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a network object")
if (is_igraph(x)){
if ( bipartite_proj){
if (is_bipartite(x)){
x <- bipartite.projection(x)[[num_proj]]
if (!is_simple(x)) x <- simplify(x)
cl <- clusters(x)
giant_comp <- induced.subgraph(x, which(cl$membership == which.max(cl$csize)))
giant_comp_edgelist <- as_edgelist(giant_comp, names = TRUE)
}
else stop("The graph is not bipartite")
}
else{
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
giant_comp <- induced.subgraph(x, which(cl$membership == which.max(cl$csize)))
giant_comp_edgelist <- as_edgelist(giant_comp, names = TRUE)
}
}
if ( is.network(x)){
edgelist<-as.edgelist(x)
x <- graph_from_edgelist(edgelist, directed = directed)
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
giant_comp <- induced.subgraph(x, which(cl$membership == which.max(cl$csize)))
giant_comp_edgelist <- as_edgelist(giant_comp, names = TRUE)
}
result <- list(giant_comp,giant_comp_edgelist)
return(result)
}
#' @title Proper centrality measure representation
#'
#' @description This function indicates the proper centrality measures of an igraph object
#' based on the network topology.
#' @param x An igraph object.
#' @details
#' This function returns a list including the names of centrality measures that are applicable
#' for the input graph based on its topology.
#' @seealso \code{\link[CINNA]{calculate_centralities}}
#'
#' @return A list including the names of centrality measures that are suitable for the input graph.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @examples
#'
#' data("zachary")
#' proper_centralities(zachary)
#'
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_directed
#' @importFrom igraph is_weighted
proper_centralities<-function(x){
if (!is_igraph(x)) stop(" Error: x must be a class of igraph object ")
if (is_directed(x) && is_weighted(x)){
proper_centralities <- c("Alpha Centrality",
"Burt's Constraint",
"Page Rank" ,
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality",
"Weighted Vertex Degree"
)
}
if (!is_directed(x) && is_weighted(x)) {
proper_centralities<-c( "subgraph centrality scores",
"Topological Coefficient",
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality",
"Weighted Vertex Degree"
)
}
if (!is_directed(x) && !is_weighted(x)) {
proper_centralities <- c( "subgraph centrality scores",
"Topological Coefficient",
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality"
)
}
if (is_directed(x) && !is_weighted(x) ) {
proper_centralities <- c("Alpha Centrality" ,
"Bonacich power centralities of positions" ,
"Page Rank" ,
"Average Distance",
"Barycenter Centrality" ,
"BottleNeck Centrality",
"Centroid value" ,
"Closeness Centrality (Freeman)" ,
"ClusterRank" ,
"Decay Centrality",
"Degree Centrality" ,
"Diffusion Degree",
"DMNC - Density of Maximum Neighborhood Component" ,
"Eccentricity Centrality" ,
"Harary Centrality",
"eigenvector centralities" ,
"K-core Decomposition" ,
"Geodesic K-Path Centrality" ,
"Katz Centrality (Katz Status Index)" ,
"Kleinberg's authority centrality scores",
"Kleinberg's hub centrality scores" ,
"clustering coefficient" ,
"Lin Centrality" ,
"Lobby Index (Centrality)" ,
"Markov Centrality" ,
"Radiality Centrality",
"Shortest-Paths Betweenness Centrality" ,
"Current-Flow Closeness Centrality",
"Closeness centrality (Latora)" ,
"Communicability Betweenness Centrality",
"Community Centrality" ,
"Cross-Clique Connectivity" ,
"Entropy Centrality" ,
"EPC - Edge Percolated Component" ,
"Laplacian Centrality" ,
"Leverage Centrality" ,
"MNC - Maximum Neighborhood Component" ,
"Hubbell Index" ,
"Semi Local Centrality",
"Closeness Vitality" ,
"Residual Closeness Centrality",
"Stress Centrality",
"Load Centrality",
"Flow Betweenness Centrality",
"Information Centrality",
"Dangalchev Closeness Centrality",
"Group Centrality",
"Harmonic Centrality",
"Local Bridging Centrality",
"Wiener Index Centrality"
)
}
print(proper_centralities)
}
#' @title Centrality measure calculation
#'
#' @description This function computes multitude centrality measures of an igraph object.
#' @param x the component of a network as an igraph object
#' @param except A vector containing names of centrality measures which could be omitted from the calculations.
#' @param include A vector including names of centrality measures which should be computed.
#' @param weights A character scalar specifying the edge attribute to use. (default=NULL)
#' @details
#' This function calculates various types of centrality measures which are applicable to the network topology
#' and returns the results as a list. In the "except" argument, you can specify centrality measures that are
#' not necessary to calculate.
#'
#' @seealso \code{\link[igraph]{alpha.centrality}}, \code{\link[igraph]{bonpow}}, \code{\link[igraph]{constraint}},
#' \code{\link[igraph]{centr_degree}}, \code{\link[igraph]{eccentricity}}, \code{\link[igraph]{eigen_centrality}},
#' \code{\link[igraph]{coreness}}, \code{\link[igraph]{authority_score}}, \code{\link[igraph]{hub_score}},
#' \code{\link[igraph]{transitivity}}, \code{\link[igraph]{page_rank}}, \code{\link[igraph]{betweenness}},
#' \code{\link[igraph]{subgraph.centrality}}, \code{\link[sna]{flowbet}}, \code{\link[sna]{infocent}},
#' \code{\link[sna]{loadcent}}, \code{\link[sna]{stresscent}}, \code{\link[sna]{graphcent}}, \code{\link[centiserve]{topocoefficient}},
#' \code{\link[centiserve]{closeness.currentflow}}, \code{\link[centiserve]{closeness.latora}},
#' \code{\link[centiserve]{communibet}}, \code{\link[centiserve]{communitycent}},
#' \code{\link[centiserve]{crossclique}}, \code{\link[centiserve]{entropy}},
#' \code{\link[centiserve]{epc}}, \code{\link[centiserve]{laplacian}}, \code{\link[centiserve]{leverage}},
#' \code{\link[centiserve]{mnc}}, \code{\link[centiserve]{hubbell}}, \code{\link[centiserve]{semilocal}},
#' \code{\link[centiserve]{closeness.vitality}},
#' \code{\link[centiserve]{closeness.residual}}, \code{\link[centiserve]{lobby}},
#' \code{\link[centiserve]{markovcent}}, \code{\link[centiserve]{radiality}}, \code{\link[centiserve]{lincent}},
#' \code{\link[centiserve]{geokpath}}, \code{\link[centiserve]{katzcent}}, \code{\link[centiserve]{diffusion.degree}},
#' \code{\link[centiserve]{dmnc}}, \code{\link[centiserve]{centroid}}, \code{\link[centiserve]{closeness.freeman}},
#' \code{\link[centiserve]{clusterrank}}, \code{\link[centiserve]{decay}},
#' \code{\link[centiserve]{barycenter}}, \code{\link[centiserve]{bottleneck}}, \code{\link[centiserve]{averagedis}},
#' \code{\link[CINNA]{local_bridging_centrality}}, \code{\link[CINNA]{wiener_index_centrality}}, \code{\link[CINNA]{group_centrality}},
#' \code{\link[CINNA]{dangalchev_closeness_centrality}}, \code{\link[CINNA]{harmonic_centrality}}, \code{\link[igraph]{strength}}
#'
#' @return A list containing centrality measure values. Each column indicates a centrality measure, and each row corresponds to a vertex.
#' The structure of the output is as follows:
#' - The list has named elements, where each element represents a centrality measure.
#' - The value of each element is a numeric vector, where each element of the vector corresponds to a vertex in the network.
#' - The order of vertices in the numeric vectors matches the order of vertices in the input igraph object.
#' - The class of the output is "centrality".
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Bonacich, P., & Lloyd, P. (2001). Eigenvector like measures of centrality for asymmetric relations. Social Networks, 23(3), 191–201.
#'
#' Bonacich, P. (1972). Factoring and weighting approaches to status scores and clique identification. The Journal of Mathematical Sociology, 2(1), 113–120.
#'
#' Bonacich, P. (1987). Power and Centrality: A Family of Measures. American Journal of Sociology, 92(5), 1170–1182.
#'
#' Burt, R. S. (2004). Structural Holes and Good Ideas. American Journal of Sociology, 110(2), 349–399.
#'
#' Batagelj, V., & Zaversnik, M. (2003). An O(m) Algorithm for Cores Decomposition of Networks, 1–9. Retrieved from
#'
#' Seidman, S. B. (1983). Network structure and minimum degree. Social Networks, 5(3), 269–287.
#'
#' Kleinberg, J. M. (1999). Authoritative sources in a hyperlinked environment. Journal of the ACM, 46(5), 604–632.
#'
#' Wasserman, S., & Faust, K. (1994). Social network analysis : methods and applications. American Ethnologist (Vol. 24).
#' Barrat, A., Barthélemy, M., Pastor Satorras, R., & Vespignani, A. (2004). The architecture of complex weighted networks. Proceedings of the National Academy of Sciences of the United States of America , 101(11), 3747–3752.
#'
#' Brin, S., & Page, L. (2010). The Anatomy of a Large Scale Hypertextual Web Search Engine The Anatomy of a Search Engine. Search, 30(June 2000), 1–7.
#'
#' Freeman, L. C. (1978). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239.
#'
#' Brandes, U. (2001). A faster algorithm for betweenness centrality*. The Journal of Mathematical Sociology, 25(2), 163–177.
#'
#' Estrada E., Rodriguez-Velazquez J. A.: Subgraph centrality in Complex Networks. Physical Review E 71, 056103.
#'
#' Freeman, L. C., Borgatti, S. P., & White, D. R. (1991). Centrality in valued graphs: A measure of betweenness based on network flow. Social Networks, 13(2), 141–154.
#'
#' Brandes, U., & Erlebach, T. (Eds.). (2005). Network Analysis (Vol. 3418). Berlin, Heidelberg: Springer Berlin Heidelberg.
#'
#' Stephenson, K., & Zelen, M. (1989). Rethinking centrality: Methods and examples. Social Networks, 11(1), 1–37.
#'
#' Wasserman, S., & Faust, K. (1994). Social network analysis : methods and applications. American Ethnologist (Vol. 24).
#'
#' Brandes, U. (2008). On variants of shortest path betweenness centrality and their generic computation. Social Networks, 30(2), 136–145.
#'
#' Goh, K.-I., Kahng, B., & Kim, D. (2001). Universal Behavior of Load Distribution in Scale Free Networks. Physical Review Letters, 87(27), 278701.
#'
#' Shimbel, A. (1953). Structural parameters of communication networks. The Bulletin of Mathematical Biophysics, 15(4), 501–507.
#'
#' Assenov, Y., Ramrez, F., Schelhorn, S.-E., Lengauer, T., & Albrecht, M. (2008). Computing topological parameters of biological networks. Bioinformatics, 24(2), 282–284.
#'
#' Diekert, V., & Durand, B. (Eds.). (2005). STACS 2005 (Vol. 3404). Berlin, Heidelberg: Springer Berlin Heidelberg.
#'
#' Gräßler, J., Koschützki, D., & Schreiber, F. (2012). CentiLib: comprehensive analysis and exploration of network centralities. Bioinformatics (Oxford, England), 28(8), 1178–9.
#'
#' Latora, V., & Marchiori, M. (2001). Efficient Behavior of Small World Networks. Physical Review Letters, 87(19), 198701.
#'
#' Opsahl, T., Agneessens, F., & Skvoretz, J. (2010). Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32(3), 245–251.
#'
#' Estrada, E., Higham, D. J., & Hatano, N. (2009). Communicability betweenness in complex networks. Physica A: Statistical Mechanics and Its Applications, 388(5), 764–774.
#'
#' Hagberg, Aric, Pieter Swart, and Daniel S Chult. Exploring network structure, dynamics, and function using NetworkX. No. LA-UR-08-05495; LA-UR-08-5495. Los Alamos National Laboratory (LANL), 2008.
#'
#' Kalinka, A. T., & Tomancak, P. (2011). linkcomm: an R package for the generation, visualization, and analysis of link communities in networks of arbitrary size and type. Bioinformatics, 27(14), 2011–2012.
#'
#' Faghani, M. R., & Nguyen, U. T. (2013). A Study of XSS Worm Propagation and Detection Mechanisms in Online Social Networks. IEEE Transactions on Information Forensics and Security, 8(11), 1815–1826.
#'
#' Brandes, U., & Erlebach, T. (Eds.). (2005). Network Analysis (Vol. 3418). Berlin, Heidelberg: Springer Berlin Heidelberg.
#'
#' Lin, C.-Y., Chin, C.-H., Wu, H.-H., Chen, S.-H., Ho, C.-W., & Ko, M.-T. (2008). Hubba: hub objects analyzer--a framework of interactome hubs identification for network biology. Nucleic Acids Research, 36(Web Server), W438–W443.
#'
#' Chin, C., Chen, S., & Wu, H. (2009). cyto Hubba: A Cytoscape Plug in for Hub Object Analysis in Network Biology. Genome Informatics …, 5(Java 5), 2–3.
#'
#' Qi, X., Fuller, E., Wu, Q., Wu, Y., & Zhang, C.-Q. (2012). Laplacian centrality: A new centrality measure for weighted networks. Information Sciences, 194, 240–253.
#'
#' Joyce, K. E., Laurienti, P. J., Burdette, J. H., & Hayasaka, S. (2010). A New Measure of Centrality for Brain Networks. PLoS ONE, 5(8), e12200.
#'
#' Lin, C.-Y., Chin, C.-H., Wu, H.-H., Chen, S.-H., Ho, C.-W., & Ko, M.-T. (2008). Hubba: hub objects analyzer--a framework of interactome hubs identification for network biology. Nucleic Acids Research, 36(Web Server), W438–W443.
#'
#' Hubbell, C. H. (1965). An Input Output Approach to Clique Identification. Sociometry, 28(4), 377.
#'
#' Dangalchev, C. (2006). Residual closeness in networks. Physica A: Statistical Mechanics and Its Applications, 365(2), 556–564.
#'
#' Brandes, U. & Erlebach, T. (2005). Network Analysis: Methodological Foundations, U.S. Government Printing Office.
#'
#' Korn, A., Schubert, A., & Telcs, A. (2009). Lobby index in networks. Physica A: Statistical Mechanics and Its Applications, 388(11), 2221–2226.
#'
#' White, S., & Smyth, P. (2003). Algorithms for estimating relative importance in networks. In Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining KDD ’03 (p. 266). New York, New York, USA: ACM Press.
#'
#' Cornish, A. J., & Markowetz, F. (2014). SANTA: Quantifying the Functional Content of Molecular Networks. PLoS Computational Biology, 10(9), e1003808.
#'
#' Scardoni, G., Petterlini, M., & Laudanna, C. (2009). Analyzing biological network parameters with CentiScaPe. Bioinformatics, 25(21), 2857–2859.
#'
#' Lin, N. (1976). Foundations of Social Research. Mcgraw Hill.
#'
#' Borgatti, S. P., & Everett, M. G. (2006). A Graph theoretic perspective on centrality. Social Networks, 28(4), 466–484.
#'
#' Newman, M. (2010). Networks. Oxford University Press.
#'
#' Junker, Bjorn H., Dirk Koschutzki, and Falk Schreiber(2006). "Exploration of biological network centralities with CentiBiN." BMC bioinformatics 7.1 : 219.
#'
#' Pal, S. K., Kundu, S., & Murthy, C. A. (2014). Centrality measures, upper bound, and influence maximization in large scale directed social networks. Fundamenta Informaticae, 130(3), 317–342.
#'
#' Lin, C.-Y., Chin, C.-H., Wu, H.-H., Chen, S.-H., Ho, C.-W., & Ko, M.-T. (2008). Hubba: hub objects analyzer--a framework of interactome hubs identification for network biology. Nucleic Acids Research, 36(Web Server), W438–W443.
#'
#' Scardoni, G., Petterlini, M., & Laudanna, C. (2009). Analyzing biological network parameters with CentiScaPe. Bioinformatics, 25(21), 2857–2859.
#'
#' Freeman, L. C. (1978). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239.
#'
#' Chen, D.-B., Gao, H., L?, L., & Zhou, T. (2013). Identifying Influential Nodes in Large Scale Directed Networks: The Role of Clustering. PLoS ONE, 8(10), e77455.
#'
#' Jana Hurajova, S. G. and T. M. (2014). Decay Centrality. In 15th Conference of Kosice Mathematicians. Herlany.
#'
#' Viswanath, M. (2009). ONTOLOGY BASED AUTOMATIC TEXT SUMMARIZATION. Vishweshwaraiah Institute of Technology.
#'
#' Przulj, N., Wigle, D. A., & Jurisica, I. (2004). Functional topology in a network of protein interactions. Bioinformatics, 20(3), 340–348.
#'
#' del Rio, G., Koschtzki, D., & Coello, G. (2009). How to identify essential genes from molecular networks BMC Systems Biology, 3(1), 102.
#'
#' Scardoni, G. and Carlo Laudanna, C.B.M.C., 2011. Network centralities for Cytoscape. University of Verona.
#'
#' BOLDI, P. & VIGNA, S. 2014. Axioms for centrality. Internet Mathematics, 00-00.
#'
#' MARCHIORI, M. & LATORA, V. 2000. Harmony in the small-world. Physica A: Statistical Mechanics and its Applications, 285, 539-546.
#'
#' OPSAHL, T., AGNEESSENS, F. & SKVORETZ, J. 2010. Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32, 245-251.
#'
#' OPSAHL, T. 2010. Closeness centrality in networks with disconnected components (http://toreopsahl.com/2010/03/20/closeness-centrality-in-networks-with-disconnected-components/)
#'
#' Michalak, T.P., Aadithya, K.V., Szczepanski, P.L., Ravindran, B. and Jennings, N.R., 2013. Efficient computation of the Shapley value for game-theoretic network centrality. Journal of Artificial Intelligence Research, 46, pp.607-650.
#'
#' Macker, J.P., 2016, November. An improved local bridging centrality model for distributed network analytics. In Military Communications Conference, MILCOM 2016-2016 IEEE (pp. 600-605). IEEE. DOI: 10.1109/MILCOM.2016.7795393
#'
#' DANGALCHEV, C. 2006. Residual closeness in networks. Physica A: Statistical Mechanics and its Applications, 365, 556-564. DOI: 10.1016/j.physa.2005.12.020
#'
#' Alain Barrat, Marc Barthelemy, Romualdo Pastor-Satorras, Alessandro Vespignani: The architecture of complex weighted networks, Proc. Natl. Acad. Sci. USA 101, 3747 (2004)
#'
#' @examples
#' data("zachary")
#' p <- proper_centralities(zachary)
#' calculate_centralities(zachary, include = "Degree Centrality")
#'
#' @export
#'
#' @importFrom igraph is_directed
#' @importFrom igraph is_connected
#' @importFrom igraph is_weighted
#' @importFrom igraph as_edgelist
#' @importFrom stats setNames
#' @importFrom igraph alpha.centrality
#' @importFrom igraph bonpow
#' @importFrom igraph constraint
#' @importFrom igraph centr_degree
#' @importFrom igraph eccentricity
#' @importFrom igraph eigen_centrality
#' @importFrom igraph coreness
#' @importFrom igraph authority_score
#' @importFrom igraph hub_score
#' @importFrom igraph transitivity
#' @importFrom igraph page_rank
#' @importFrom igraph betweenness
#' @importFrom igraph subgraph.centrality
#' @importFrom igraph strength
#' @importFrom network network
#' @importFrom sna flowbet
#' @importFrom sna infocent
#' @importFrom sna loadcent
#' @importFrom sna stresscent
#' @importFrom sna graphcent
#' @importFrom centiserve topocoefficient
#' @importFrom centiserve closeness.currentflow
#' @importFrom centiserve closeness.latora
#' @importFrom centiserve communibet
#' @importFrom centiserve communitycent
#' @importFrom centiserve crossclique
#' @importFrom centiserve entropy
#' @importFrom centiserve epc
#' @importFrom centiserve laplacian
#' @importFrom centiserve leverage
#' @importFrom centiserve mnc
#' @importFrom centiserve hubbell
#' @importFrom centiserve semilocal
#' @importFrom centiserve closeness.vitality
#' @importFrom centiserve closeness.residual
#' @importFrom centiserve lobby
#' @importFrom centiserve markovcent
#' @importFrom centiserve radiality
#' @importFrom centiserve lincent
#' @importFrom centiserve geokpath
#' @importFrom centiserve katzcent
#' @importFrom centiserve diffusion.degree
#' @importFrom centiserve dmnc
#' @importFrom centiserve centroid
#' @importFrom centiserve closeness.freeman
#' @importFrom centiserve clusterrank
#' @importFrom centiserve decay
#' @importFrom centiserve barycenter
#' @importFrom centiserve bottleneck
#' @importFrom centiserve averagedis
calculate_centralities <- function( x, except = NULL, include = NULL, weights = NULL){
if (!("igraph" %in% class(x) && is_connected(x))) stop("The input is not an igraph object
or may not be connected.")
y <- as_edgelist(x)
y <- network(y)
if (is_directed(x) && is_weighted(x) ){
centrality_funcs <- list(
"Alpha Centrality" = function(x) alpha.centrality(x, weights = weights),
"Burt's Constraint" = function(x)constraint(x, weights = weights),
"Page Rank" = function(x)page_rank(x)$vector,
"Average Distance" = function(x)averagedis(x, weights = weights),
"Barycenter Centrality" = function(x)barycenter(x, weights = weights),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = weights),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = weights),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = weights),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = weights)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = weights),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = weights)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = weights)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = weights,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = weights),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = weights),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = weights),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = weights),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = weights),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = weights),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = weights),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = weights),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = weights),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = weights),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = weights),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = weights)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
if (!is_directed(x) && is_weighted(x)){
centrality_funcs <- list(
"subgraph centrality scores" = function(x)subgraph.centrality(x),
"Topological Coefficient" = function(x)topocoefficient(x),
"Average Distance" = function(x)averagedis(x, weights = weights),
"Barycenter Centrality" = function(x)barycenter(x, weights = weights),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = weights),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = weights),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = weights),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = weights)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = weights),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = weights)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = weights)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = weights,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = weights),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = weights),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = weights),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = weights),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = weights),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = weights),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = weights),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = weights),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = weights),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="undirected"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = weights),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = weights),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = weights)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
if(!is_directed(x) && !is_weighted(x)) {
centrality_funcs <- list(
"subgraph centrality scores" = function(x)subgraph.centrality(x),
"Topological Coefficient" = function(x)topocoefficient(x),
"Average Distance" = function(x)averagedis(x, weights = NULL),
"Barycenter Centrality" = function(x)barycenter(x, weights = NULL),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = NULL),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = NULL),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = NULL),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = NULL)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = NULL),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = NULL)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = NULL)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = NULL,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = NULL),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = NULL),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = NULL),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = NULL),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = NULL),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = NULL),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = NULL),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = NULL),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
if(is_directed(x) && !is_weighted(x) ) {
centrality_funcs <- list(
"Alpha Centrality" = function(x) alpha.centrality(x, weights = NULL),
"Bonacich power centralities of positions" = function(x)bonpow(x),
"Page Rank" = function(x)page_rank(x)$vector,
"Average Distance" = function(x)averagedis(x, weights = NULL),
"Barycenter Centrality" = function(x)barycenter(x, weights = NULL),
"BottleNeck Centrality" = function(x)bottleneck(x),
"Centroid value" = function(x)centroid(x, weights = NULL),
"Closeness Centrality (Freeman)" = function(x)closeness.freeman(x, weights = NULL),
"ClusterRank" = function(x)clusterrank(x),
"Decay Centrality" = function(x)decay(x, weights = NULL),
"Degree Centrality" = function(x)centr_degree(x)$res,
"Diffusion Degree" = function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component" = function(x)dmnc(x),
"Eccentricity Centrality" = function(x)eccentricity(x),
"eigenvector centralities" = function(x)eigen_centrality(x, weights = NULL)$vector,
"K-core Decomposition" = function(x)coreness(x),
"Geodesic K-Path Centrality" = function(x)geokpath(x, weights = NULL),
"Katz Centrality (Katz Status Index)" = function(x)katzcent(x),
"Kleinberg's authority centrality scores" = function(x)authority_score(x, weights = NULL)$vector,
"Kleinberg's hub centrality scores" = function(x)hub_score(x, weights = NULL)$vector,
"clustering coefficient" = function(x)transitivity(x, weights = NULL,type="local"),
"Lin Centrality" = function(x)lincent(x, weights = NULL),
"Lobby Index (Centrality)" = function(x)lobby(x),
"Markov Centrality" = function(x)markovcent(x),
"Radiality Centrality" = function(x)radiality(x, weights = NULL),
"Shortest-Paths Betweenness Centrality" = function(x)betweenness(x),
"Current-Flow Closeness Centrality" = function(x)closeness.currentflow(x, weights = NULL),
"Closeness centrality (Latora)" = function(x)closeness.latora(x, weights = NULL),
"Communicability Betweenness Centrality" = function(x)communibet(x),
"Community Centrality" = function(x)communitycent(x),
"Cross-Clique Connectivity" = function(x)crossclique(x),
"Entropy Centrality" = function(x)entropy(x, weights = NULL),
"EPC - Edge Percolated Component" = function(x)epc(x),
"Laplacian Centrality" = function(x)laplacian(x),
"Leverage Centrality" = function(x)leverage(x),
"MNC - Maximum Neighborhood Component" = function(x)mnc(x),
"Hubbell Index" = function(x)hubbell(x, weights = NULL),
"Semi Local Centrality" = function(x)semilocal(x),
"Closeness Vitality" = function(x)closeness.vitality(x, weights = NULL),
"Residual Closeness Centrality" = function(x)closeness.residual(x, weights = NULL),
"Stress Centrality" = function(x)stresscent(y),
"Load Centrality" = function(x)loadcent(y),
"Flow Betweenness Centrality" = function(x)flowbet(y),
"Information Centrality" = function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL)
)
if (!is.null(include)){
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), include)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
else{
centrality_funcs <- centrality_funcs[setdiff(names(centrality_funcs), except)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
return(result)
}
}
}
#' @title PCA Centrality Measures
#'
#' @description This function performs Principal Component Analysis (PCA) on centrality measures.
#' It computes the contributions of variables to the principal components and visualizes them.
#'
#' @param x a list containing the computed centrality values
#' @param scale.unit a boolean constant indicating whether data should be scaled to unit variance (default = TRUE)
#' @param cut.off The intensity that must be exceeded in cumulative percentage of variance of eigen values (default = 80)
#' @param ncp number of dimensions in the final results (default = 5)
#' @param graph a boolean constant indicating whether the graph should be displayed (default = FALSE)
#' @param axes a length 2 vector describing the number of components to plot (default = c(1, 2))
#'
#' @return A plot illustrating the contributions of variables to the principal components.
#' The x-axis represents the centrality measures, and the y-axis represents the contribution.
#' The higher the contribution value, the more important the centrality measure is in the ranking.
#' The plot helps in identifying the most influential centrality measures.
#'
#' @importFrom FactoMineR PCA
#' @importFrom factoextra fviz_contrib
#' @importFrom ggplot2 labs theme_grey theme element_text element_rect
#' @importFrom stats na.omit
#'
#' @examples
#' # Create a data frame with multiple observations
#' centralities <- data.frame(
#' Betweenness = c(0.2, 0.3, 0.5),
#' Closeness = c(0.4, 0.2, 0.6),
#' Degree = c(0.3, 0.1, 0.4),
#' Eigenvector = c(0.1, 0.5, 0.2)
#' )
#' pca_centralities(centralities)
#'
#' @export
pca_centralities <- function(x, scale.unit = TRUE, cut.off = 80, ncp = 2, graph = FALSE, axes = c(1, 2)) {
requireNamespace("FactoMineR", quietly = TRUE)
requireNamespace("factoextra", quietly = TRUE)
requireNamespace("ggplot2", quietly = TRUE)
requireNamespace("utils", quietly = TRUE)
x <- x[!sapply(x, is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
res_pca <- FactoMineR::PCA(x, scale.unit = scale.unit, ncp = ncp, graph = graph, axes = axes)
PCs <- table(res_pca$eig[,"cumulative percentage of variance"] > cut.off)["FALSE"] + 1
if (is.na(PCs)) PCs <- 1
factoextra::fviz_contrib(res_pca, choice = "var", axes = 1:PCs, color = "black", fill = "turquoise") +
ggplot2::labs(x = "\nCentrality measures", y = "Contributions\n") +
ggplot2::theme_grey() +
ggplot2::ggtitle("Contribution of variables via PCA") +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle = 45, vjust = 0.5),
plot.title = ggplot2::element_text(hjust = 0.5),
# Adjust line size here
line = ggplot2::element_line(color = "black", size = 1.5)
)
}
#' @title Graph visualization based on a specific centrality measure
#'
#' @description This function demonstrates the input graph in which the size of nodes
#' indicates calculated centrality value.
#' @param x an igraph object
#' @param computed_centrality_value A vector containing the values of calculated centrality measure for each node.
#' @param centrality.type The type of centrality which should be calculated.
#' @details
#' This function represents the graph in which size of nodes are based on computed centrality value. If the values of wanted centrality
#' measure were computed then by placing them in computed_centrality_value argument to use it for drawing the plot. Otherwise, by only giving the
#' name of favorite centrality measure in centrality.type argument, this function will calculate it and then demonstrates the corresponding graph.
#' @return A plot illustrating the input graph, where the size of nodes represents the calculated centrality values.
#' The function takes an igraph object as input and either a vector of precomputed centrality values or the name of the desired centrality measure.
#' If precomputed centrality values are provided in the computed_centrality_value argument, the graph will be plotted based on those values.
#' If only the name of the desired centrality measure is provided in the centrality.type argument, the function will calculate the centrality values
#' and then plot the graph.
#' The plot shows the nodes of the graph, with their sizes indicating the centrality values. The larger the node, the higher the centrality value.
#' This plot helps visualize the distribution of centrality values across the nodes of the graph.
#' The function returns the plot and does not modify the input graph.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#'
#' @importFrom igraph is.directed
#' @importFrom igraph alpha.centrality
#' @importFrom igraph bonpow
#' @importFrom igraph constraint
#' @importFrom igraph centr_degree
#' @importFrom igraph eccentricity
#' @importFrom igraph eigen_centrality
#' @importFrom igraph coreness
#' @importFrom igraph authority_score
#' @importFrom igraph hub_score
#' @importFrom igraph transitivity
#' @importFrom igraph page_rank
#' @importFrom igraph betweenness
#' @importFrom igraph subgraph.centrality
#' @importFrom igraph strength
#' @importFrom sna flowbet
#' @importFrom sna infocent
#' @importFrom sna loadcent
#' @importFrom sna stresscent
#' @importFrom sna graphcent
#' @importFrom centiserve topocoefficient
#' @importFrom centiserve closeness.currentflow
#' @importFrom centiserve closeness.latora
#' @importFrom centiserve communibet
#' @importFrom centiserve communitycent
#' @importFrom centiserve crossclique
#' @importFrom centiserve entropy
#' @importFrom centiserve epc
#' @importFrom centiserve laplacian
#' @importFrom centiserve leverage
#' @importFrom centiserve mnc
#' @importFrom centiserve hubbell
#' @importFrom centiserve semilocal
#' @importFrom centiserve closeness.vitality
#' @importFrom centiserve closeness.residual
#' @importFrom centiserve lobby
#' @importFrom centiserve markovcent
#' @importFrom centiserve radiality
#' @importFrom centiserve lincent
#' @importFrom centiserve geokpath
#' @importFrom centiserve katzcent
#' @importFrom centiserve diffusion.degree
#' @importFrom centiserve dmnc
#' @importFrom centiserve centroid
#' @importFrom centiserve closeness.freeman
#' @importFrom centiserve clusterrank
#' @importFrom centiserve decay
#' @importFrom centiserve barycenter
#' @importFrom centiserve bottleneck
#' @importFrom centiserve averagedis
#' @importFrom igraph layout_in_circle
#' @importFrom graphics plot
#' @importFrom igraph strength
visualize_graph <- function( x , computed_centrality_value=NULL , centrality.type="Degree Centrality"){
if (is.null(computed_centrality_value)){
if (!("igraph" %in% class(x) && is_connected(x) )) stop("The input is not an igraph object
or may not be connected.")
y <- as_edgelist(x)
y <- network(y)
centrality_funcs <- list(
"subgraph centrality scores"=function(x)subgraph.centrality(x),
"Topological Coefficient"=function(x)topocoefficient(x),
"Alpha Centrality"=function(x) alpha.centrality(x),
"Bonacich power centralities of positions"=function(x)bonpow(x),
"Burt's Constraint"=function(x)constraint(x),
"Page Rank"=function(x)page_rank(x)$vector,
"Average Distance"=function(x)averagedis(x),
"Barycenter Centrality"=function(x)barycenter(x),
"BottleNeck Centrality"=function(x)bottleneck(x),
"Centroid value"=function(x)centroid(x),
"Closeness Centrality (Freeman)"=function(x)closeness.freeman(x),
"ClusterRank"=function(x)clusterrank(x),
"Decay Centrality"=function(x)decay(x),
"Degree Centrality"=function(x)centr_degree(x)$res,
"Diffusion Degree"=function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component"=function(x)dmnc(x),
"Eccentricity Centrality"=function(x)eccentricity(x),
"eigenvector centralities"=function(x)eigen_centrality(x)$vector,
"K-core Decomposition"=function(x)coreness(x),
"Geodesic K-Path Centrality"=function(x)geokpath(x),
"Katz Centrality (Katz Status Index)"=function(x)katzcent(x),
"Kleinberg's authority centrality scores"=function(x)authority_score(x)$vector,
"Kleinberg's hub centrality scores"=function(x)hub_score(x)$vector,
"clustering coefficient"=function(x)transitivity(x,type="local"),
"Lin Centrality"=function(x)lincent(x),
"Lobby Index (Centrality)"=function(x)lobby(x),
"Markov Centrality"=function(x)markovcent(x),
"Radiality Centrality"=function(x)radiality(x),
"Shortest-Paths Betweenness Centrality"=function(x)betweenness(x),
"Current-Flow Closeness Centrality"=function(x)closeness.currentflow(x),
"Closeness centrality (Latora)"=function(x)closeness.latora(x),
"Communicability Betweenness Centrality"=function(x)communibet(x),
"Community Centrality"=function(x)communitycent(x),
"Cross-Clique Connectivity"=function(x)crossclique(x),
"Entropy Centrality"=function(x)entropy(x),
"EPC - Edge Percolated Component"=function(x)epc(x),
"Laplacian Centrality"=function(x)laplacian(x),
"Leverage Centrality"=function(x)leverage(x),
"MNC - Maximum Neighborhood Component"=function(x)mnc(x),
"Hubbell Index"=function(x)hubbell(x),
"Semi Local Centrality"=function(x)semilocal(x),
"Closeness Vitality"=function(x)closeness.vitality(x),
"Residual Closeness Centrality"=function(x)closeness.residual(x),
"Stress Centrality"=function(x)stresscent(y),
"Load Centrality"=function(x)loadcent(y),
"Flow Betweenness Centrality"=function(x)flowbet(y),
"Information Centrality"=function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = NULL)
)
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), centrality.type)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
result <- result[!sapply(result,is.null)]
result <- as.data.frame(result)
result <- scale(result, center = FALSE, scale = TRUE)
plot(x, vertex.size=result*10, layout= layout_in_circle(x,order(result*10) ) ,
vertex.color="turquoise",vertex.frame.color="orange2")
}
else{
computed_centrality_value<-scale(computed_centrality_value, center = FALSE, scale = TRUE)
plot(x, vertex.size=computed_centrality_value*10,
layout= layout_in_circle(x,order(computed_centrality_value*10) ) ,
vertex.color="turquoise",vertex.frame.color="orange2")
}
}
#' @title Pairwise association plot between centrality measures
#'
#' @description This function computes regression between pair of centrality measures
#' to show more details of association among them.
#' @param x a vector containing a centrality measure as independent variable
#' @param y a vector containing a centrality measure as dependent variable
#' @param scale Whether the centrality values should be scaled or not
#' @details
#' This function applies regression analysis on two different centrality values in order to find out the corresponding association
#' between them. Regression analysis is a kind of statitiscal method for approximation the association
#' between variables.It asserts that the value of dependent variable changes when the
#' value of independent variable varies.
#' @return The function returns a regression plot and the results of the regression analysis between two centrality measures.
#' The function takes two vectors, `x` and `y`, which represent the independent and dependent centrality measures, respectively.
#' If the `scale` parameter is set to `TRUE`, the centrality values will be scaled before performing the regression analysis.
#' The function applies regression analysis to estimate the association between the two centrality measures.
#' It creates a scatter plot of the centrality values, with a regression line representing the relationship between the variables.
#' Additionally, the plot includes confidence intervals around the regression line to show the uncertainty of the estimates.
#' The function returns the regression plot and the computed regression results.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Chambers, J. M. (1992). Statistical Models in S. Wadsworth. Pacific Grove, California. Retrieved from
#' [Link to the reference]
#'
#' Wilkinson, G. N., & Rogers, C. E. (1973). Symbolic Description of Factorial Models for Analysis of Variance. Applied Statistics, 22(3), 392.
#'
#' @export
#'
#' @importFrom ggplot2 ggplot
#' @importFrom ggplot2 stat_summary
#' @importFrom ggplot2 geom_smooth
#' @importFrom ggplot2 mean_cl_normal
#' @importFrom ggplot2 aes
#' @importFrom ggplot2 xlab
#' @importFrom ggplot2 ylab
#' @importFrom ggplot2 geom_point
#' @importFrom stats lm
visualize_association <- function( x , y, scale=TRUE){
xname <- substitute(x)
yname <- substitute(y)
df <- as.data.frame(cbind(x,y))
names(df) <- c(xname, yname)
if (scale%in%TRUE){
df <- scale(df, center = TRUE, scale = scale)
df <- as.data.frame(df)
visualization <- ggplot(data=df,aes(x,y)) +
stat_summary(fun.data=mean_cl_normal,geom="point",fill="skyblue3",shape=21,colour="black", size = 3) +
geom_smooth(method='lm',colour = 'red') + xlab(xname) +
ylab(yname)+geom_point(shape=21, fill="blue", color="darkred", size=3)
linear.regression <- lm(df[,2]~df[,1])
return (list(linear.regression=linear.regression, visualization=visualization))
}
else {visualization <- ggplot(data=df,aes(x,y)) +
stat_summary(fun.data=mean_cl_normal,geom="point",fill="skyblue3",shape=21,colour="black", size = 3) +
geom_smooth(method='lm',colour = 'red') + xlab(xname) +
ylab(yname)+geom_point(shape=21, fill="blue", color="darkred", size=3)
linear.regression <- lm(y~x)
return (list(linear.regression=linear.regression, visualization=visualization))}
}
#' @title Pairwise correlation plot between two centrality measures
#'
#' @description This function computes and plots correlation between pair of centrality
#' measures and histogram plots.
#' @param x a vector containing a centrality measure
#' @param y a vector containing another centrality measure
#' @param scale Whether the centrality values should be scaled or not
#' @details
#' This function illustrates the correlation value between two centrality measures and their
#' corresponding scatterplot and histograms.
#' @seealso \code{\link[GGally]{ggpairs}}
#' @return The function computes and plots the correlation between two centrality measures.
#' The function takes two vectors, `x` and `y`, representing the centrality measures to be correlated.
#' If the `scale` parameter is set to `TRUE`, the centrality values will be scaled before computing the correlation.
#' The function creates a correlation plot, which includes a scatterplot of the centrality values and histograms of each measure.
#' The correlation coefficient between the two measures is displayed on the plot.
#' The function returns the correlation plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Emerson, J. W., Green, W. A., Schloerke, B., Crowley, J., Cook, D., Hofmann, H., & Wickham, H. (2013). The Generalized Pairs Plot. Journal of Computational and Graphical Statistics, 22(1), 79–91.
#'
#' @export
#'
#' @importFrom GGally ggpairs
visualize_pair_correlation <- function( x , y, scale=TRUE){
xname <- substitute(x)
yname <- substitute(y)
df <- as.data.frame(cbind(x,y))
names(df) <- c(xname, yname)
if (scale%in%TRUE){
df <- scale(df, center = TRUE, scale = scale)
df <- as.data.frame(df)
ggpairs(df, columns=1:2)
}
else ggpairs(df, columns=1:2)
}
#' @title Heatmap plot between centrality measures
#'
#' @description This function draws heatmap between pair of centrality measures
#' @param x a list indicating calculated centrality measures
#' @param scale Whether the centrality values should be scaled or not
#' @details
#' This function illustrates the heatmap plot of computed centrality measures.
#' @return The function generates a heatmap plot between pairs of centrality measures.
#' The function takes a list x as input, which contains the computed centrality measures.
#' If the scale parameter is set to TRUE, the centrality values will be scaled before plotting the heatmap.
#' The function creates a heatmap plot, where each cell represents the correlation between a pair of centrality measures.
#' The color of each cell indicates the strength and direction of the correlation.
#' The function returns the heatmap plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#'
#' @importFrom pheatmap pheatmap
visualize_heatmap <- function( x, scale = TRUE ){
if(scale%in%TRUE){
x <- x[!sapply(x,is.null)]
d <- as.data.frame(x)
x <- scale(d, center = TRUE, scale = scale)
rownames(x)<- rownames(d)
pheatmap(x, legend = TRUE, cluster_rows = TRUE, cluster_cols = TRUE, fontsize = 8)
}
else{
pheatmap(x, legend = TRUE, cluster_rows = TRUE, cluster_cols = TRUE, fontsize = 8)
}
}
#' @title Correlation plot between centrality measures
#'
#' @description This function draw correlation plot between pair of centrality measures
#' @param x a list indicating calculated centrality measures
#' @param scale Whether the centrality values should be scaled or not(default=TRUE)
#' @param method a character string describing the type of correlation coefficient (or covariance)
#' to be computed. The proper values are "pearson", "kendall", or "spearman". (default="pearson")
#' @details
#' This function illustrates pairwise correlation plot of computed centrality measures.
#' The names of centralities shown in the result plot is abbreviated and compelete names
#' can be seen in "proper_centralities" function.
#' Colors from red to blue indicate the intensity of correlation value. If two centrality
#' measures have an inverse relationship then their correspnding color in plot have to be red
#' and vice versa.
#'
#' @seealso \code{\link[GGally]{ggpairs}}
#' @return The function generates a correlation plot between pairs of centrality measures.
#' The function takes a list x as input, which contains the computed centrality measures.
#' If the scale parameter is set to TRUE, the centrality values will be scaled before computing the correlations.
#' The method parameter specifies the type of correlation coefficient to be computed: "pearson" (default), "kendall", or "spearman".
#' The function creates a pairwise correlation plot, where each cell represents the correlation between a pair of centrality measures.
#' The color of each cell indicates the strength and direction of the correlation, ranging from red (negative correlation) to blue (positive correlation).
#' The function returns the pairwise correlation plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @export
#'
#' @importFrom corrplot corrplot.mixed
#' @importFrom stats cor
visualize_correlations <- function(x, scale=TRUE,method = "pearson"){
if (scale%in%TRUE){
x <- x[!sapply(x,is.null)]
d <- as.data.frame(x)
x <- scale(d, center = TRUE, scale = scale)
name <- unlist(lapply(strsplit(colnames(x), '.', fixed = TRUE), '[', 1))
colnames(x) <- name
M <- cor(x,method = method)
corrplot.mixed(M,tl.cex =0.75)
}
else{
x <- x[!sapply(x,is.null)]
d <- as.data.frame(x)
x <- as.matrix(d)
name <- unlist(lapply(strsplit(colnames(x), '.', fixed = TRUE), '[', 1))
colnames(x) <- name
M <- cor(x,method = method)
corrplot.mixed(M,tl.cex =0.75)
}
}
#' @title Dendrogram plot among centrality measures
#'
#' @description This function demonstrates the vertice dendrogram of a graph based
#' on a centrality type.
#' @param x an igraph object
#' @param centrality.type The type of centrality which should be considered.(default="Degree Centrality")
#' @param computed_centrality_value A vector containing the values of calculated centrality measure for each node.(default=NULL)
#' @param k number of clusters(default=4)
#' @return The function generates a dendrogram plot of a graph's vertices based on a centrality measure.
#' The function takes an igraph object `x` as input, representing the graph.
#' The `centrality.type` parameter specifies the type of centrality measure to be considered (default is "Degree Centrality").
#' If the `computed_centrality_value` parameter is provided, it should be a vector containing the computed centrality values for each node.
#' Otherwise, the function will compute the specified centrality measure for the graph.
#' The `k` parameter specifies the number of clusters to be formed from the dendrogram (default is 4).
#' The function creates a dendrogram plot of the vertices, where the branching structure represents the hierarchical clustering based on the centrality measure.
#' The function returns the dendrogram plot.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Galili, T. (2015). dendextend: an R package for visualizing, adjusting and comparing trees of hierarchical clustering. Bioinformatics, 31(22), 3718–3720.
#'
#' @export
#'
#' @importFrom igraph alpha.centrality
#' @importFrom igraph bonpow
#' @importFrom igraph constraint
#' @importFrom igraph centr_degree
#' @importFrom igraph eccentricity
#' @importFrom igraph eigen_centrality
#' @importFrom igraph coreness
#' @importFrom igraph authority_score
#' @importFrom igraph hub_score
#' @importFrom igraph transitivity
#' @importFrom igraph page_rank
#' @importFrom igraph betweenness
#' @importFrom igraph subgraph.centrality
#' @importFrom igraph strength
#' @importFrom sna flowbet
#' @importFrom sna infocent
#' @importFrom sna loadcent
#' @importFrom sna stresscent
#' @importFrom sna graphcent
#' @importFrom centiserve topocoefficient
#' @importFrom centiserve closeness.currentflow
#' @importFrom centiserve closeness.latora
#' @importFrom centiserve communibet
#' @importFrom centiserve communitycent
#' @importFrom centiserve crossclique
#' @importFrom centiserve entropy
#' @importFrom centiserve epc
#' @importFrom centiserve laplacian
#' @importFrom centiserve leverage
#' @importFrom centiserve mnc
#' @importFrom centiserve hubbell
#' @importFrom centiserve semilocal
#' @importFrom centiserve closeness.vitality
#' @importFrom centiserve closeness.residual
#' @importFrom centiserve lobby
#' @importFrom centiserve markovcent
#' @importFrom centiserve radiality
#' @importFrom centiserve lincent
#' @importFrom centiserve geokpath
#' @importFrom centiserve katzcent
#' @importFrom centiserve diffusion.degree
#' @importFrom centiserve dmnc
#' @importFrom centiserve centroid
#' @importFrom centiserve closeness.freeman
#' @importFrom centiserve clusterrank
#' @importFrom centiserve decay
#' @importFrom centiserve barycenter
#' @importFrom centiserve bottleneck
#' @importFrom centiserve averagedis
#' @importFrom dendextend circlize_dendrogram
#' @importFrom dendextend set
#' @importFrom dendextend highlight_branches_col
#' @importFrom viridis viridis
#' @importFrom stats dist
#' @importFrom stats hclust
#' @importFrom stats as.dendrogram
visualize_dendrogram <- function( x, centrality.type="Degree Centrality", computed_centrality_value=NULL , k=4){
if (is.null(computed_centrality_value)){
if (!("igraph" %in% class(x) && is_connected(x) )) stop("The input is not an igraph object
or may not be connected.")
y <- as_edgelist(x)
y <- network(y)
centrality_funcs <- list(
"subgraph centrality scores"=function(x)subgraph.centrality(x),
"Topological Coefficient"=function(x)topocoefficient(x),
"Alpha Centrality"=function(x) alpha.centrality(x),
"Bonacich power centralities of positions"=function(x)bonpow(x),
"Burt's Constraint"=function(x)constraint(x),
"Page Rank"=function(x)page_rank(x)$vector,
"Average Distance"=function(x)averagedis(x),
"Barycenter Centrality"=function(x)barycenter(x),
"BottleNeck Centrality"=function(x)bottleneck(x),
"Centroid value"=function(x)centroid(x),
"Closeness Centrality (Freeman)"=function(x)closeness.freeman(x),
"ClusterRank"=function(x)clusterrank(x),
"Decay Centrality"=function(x)decay(x),
"Degree Centrality"=function(x)centr_degree(x)$res,
"Diffusion Degree"=function(x)diffusion.degree(x),
"DMNC - Density of Maximum Neighborhood Component"=function(x)dmnc(x),
"Eccentricity Centrality"=function(x)eccentricity(x),
"eigenvector centralities"=function(x)eigen_centrality(x)$vector,
"K-core Decomposition"=function(x)coreness(x),
"Geodesic K-Path Centrality"=function(x)geokpath(x),
"Katz Centrality (Katz Status Index)"=function(x)katzcent(x),
"Kleinberg's authority centrality scores"=function(x)authority_score(x)$vector,
"Kleinberg's hub centrality scores"=function(x)hub_score(x)$vector,
"clustering coefficient"=function(x)transitivity(x,type="local"),
"Lin Centrality"=function(x)lincent(x),
"Lobby Index (Centrality)"=function(x)lobby(x),
"Markov Centrality"=function(x)markovcent(x),
"Radiality Centrality"=function(x)radiality(x),
"Shortest-Paths Betweenness Centrality"=function(x)betweenness(x),
"Current-Flow Closeness Centrality"=function(x)closeness.currentflow(x),
"Closeness centrality (Latora)"=function(x)closeness.latora(x),
"Communicability Betweenness Centrality"=function(x)communibet(x),
"Community Centrality"=function(x)communitycent(x),
"Cross-Clique Connectivity"=function(x)crossclique(x),
"Entropy Centrality"=function(x)entropy(x),
"EPC - Edge Percolated Component"=function(x)epc(x),
"Laplacian Centrality"=function(x)laplacian(x),
"Leverage Centrality"=function(x)leverage(x),
"MNC - Maximum Neighborhood Component"=function(x)mnc(x),
"Hubbell Index"=function(x)hubbell(x),
"Semi Local Centrality"=function(x)semilocal(x),
"Closeness Vitality"=function(x)closeness.vitality(x),
"Residual Closeness Centrality"=function(x)closeness.residual(x),
"Stress Centrality"=function(x)stresscent(y),
"Load Centrality"=function(x)loadcent(y),
"Flow Betweenness Centrality"=function(x)flowbet(y),
"Information Centrality"=function(x)infocent(y),
"Harary Centrality" = function(x)graphcent(y, gmode="graph", diag=T, cmode="directed"),
"Dangalchev Closeness Centrality"= function(x)dangalchev_closeness_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Group Centrality"= function(x)group_centrality(x, vids = V(x)),
"Harmonic Centrality"= function(x)harmonic_centrality(x, vids = V(x), mode = "all", weights = NULL),
"Local Bridging Centrality"= function(x)local_bridging_centrality(x, vids = V(x)),
"Wiener Index Centrality"= function(x)wiener_index_centrality(x, vids = V(x), mode ="all", weights = NULL),
"Weighted Vertex Degree" = function(x)strength(x, vids = V(x), mode ="all", weights = NULL)
)
centrality_funcs <- centrality_funcs[intersect(names(centrality_funcs), centrality.type)]
n <- names(centrality_funcs)
warningsText <- ""
result <- lapply(setNames(n, n),
function(functionName, x) {
f <- centrality_funcs[[functionName]]
tryCatch(f(x),
error = function(e) {
warningsText <- paste0(warningsText,
"\nError in ", functionName, ":\n", e$message)
return(NULL)
})
}, x)
if (nchar(warningsText) > 0)
warning(warningsText)
result <-result[!sapply(result,is.null)]
result <- as.data.frame(result)
rownames(result) = V(x)$name
result <- scale(result, center = TRUE, scale = TRUE)
dend <- result%>% dist %>% hclust %>% as.dendrogram %>%
highlight_branches_col(viridis(100)) %>%
set("branches_k_color", k=3)%>%
set("labels_colors")%>%set("nodes_pch", 20)
circlize_dendrogram(dend, labels_track_height = NA, dend_track_height = .4)
}
else{
computed_centrality_value <- scale(computed_centrality_value, center = FALSE, scale = TRUE)
names(computed_centrality_value) = V(x)$name
dend <- computed_centrality_value%>% dist %>% hclust %>% as.dendrogram %>%
highlight_branches_col(viridis(100)) %>% set("branches_k_color", k=3)%>%
set("labels_colors")%>%set("nodes_pch", 20)
circlize_dendrogram(dend, labels_track_height = NA, dend_track_height = .4)
}
}
#' @title Summarize PCA result related to centrality measures
#'
#' @description This function summarizes the PCA result related to centrality measures.
#' @param x A list containing the computed centrality values.
#' @param scale.unit A boolean value indicating whether the data should be scaled to unit variance (default = TRUE).
#' @param ncp The number of dimensions in the final results (default = 5).
#' @return The result of the \code{pca_centralities} function, which includes the PCA analysis results such as eigenvalues, variance explained, and scores.
#' The returned value is an object of class "PCA" from the FactoMineR package. It contains the following components:
#' \item{eig}{A numeric vector of eigenvalues, indicating the amount of variance explained by each principal component.}
#' \item{var}{A numeric vector of proportions of variance explained by each principal component.}
#' \item{ind}{A data frame of individual scores, where each row represents an individual and each column represents a principal component.}
#' \item{call}{The function call used to create the PCA object.}
#' \item{call2}{The call used to compute the PCA analysis.}
#' \item{call3}{The call used to project the individuals.}
#' \item{svd}{The singular value decomposition of the data matrix.}
#' \item{cor}{The correlation matrix of the variables.}
#' \item{cos2}{The squared cosines of the variables.}
#' \item{contrib}{The contributions of the variables to the principal components.}
#' \item{proj}{The projected coordinates of the individuals on the principal components.}
#' \item{quali.sup}{The results of the supplementary qualitative variables analysis.}
#' \item{quali.sup.ind}{The results of the supplementary individuals analysis.}
#' \item{quanti.sup}{The results of the supplementary quantitative variables analysis.}
#' \item{quanti.sup.var}{The results of the supplementary variables analysis.}
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom FactoMineR PCA
#' @importFrom stats na.omit
summary_pca_centralities <- function( x , scale.unit = TRUE,ncp = 5){
x <- x[!sapply(x,is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
res_pca <- PCA(x, scale.unit = scale.unit, ncp = ncp, graph = FALSE)
l <- list(eigen= res_pca$eig, contribution= res_pca$var$contrib)
print(l)
}
#' @title Summarize component extraction of a graph
#'
#' @description This function summarizes all components of the input which can be an "igraph" object or a "network" object.
#' @param x An igraph or a network object.
#' @param directed A boolean value indicating whether to create a directed graph (default = TRUE).
#' @param bipartite_proj A boolean value indicating whether the bipartite network should be projected (default = FALSE).
#' @param num_proj A number indicating the number of projects specifically for bipartite graphs (default = 1).
#' @return A list of igraph objects representing the extracted components of the graph.
#' Each element in the list corresponds to a component, and each component is represented by an igraph object.
#' The components are extracted based on the connectivity of the graph and can include single nodes, disconnected subgraphs, or connected subgraphs.
#' If the graph is directed and the \code{directed} parameter is set to FALSE, the function returns the weakly connected components.
#' If the graph is bipartite and the \code{bipartite_proj} parameter is set to TRUE, the function returns the projected graph components.
#' If the graph is neither directed nor bipartite, the function returns the connected components.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom igraph is_igraph
#' @importFrom igraph is_bipartite
#' @importFrom igraph bipartite.projection
#' @importFrom igraph is_simple
#' @importFrom igraph simplify
#' @importFrom igraph clusters
#' @importFrom igraph induced.subgraph
#' @importFrom igraph graph_from_edgelist
#' @importFrom network is.network
#' @importFrom network as.edgelist
#' @importFrom network is.network
#' @importFrom network network
#' @importFrom stats na.omit
summary_graph_extract_components <- function( x, directed = TRUE, bipartite_proj = FALSE , num_proj = 1){
if (!("igraph" %in% class(x) || ("network" %in% class(x) ))) stop("The input is not an igraph or a network object")
if (is_igraph(x)) {
if (bipartite_proj){
if (is_bipartite(x)){
x <- bipartite.projection(x)[[num_proj]]
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
}
else{
if (!is_simple(x)) x<-simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
}
if( is.network(x)){
edgelist <- as.edgelist(x)
x <- graph_from_edgelist(edgelist, directed = directed)
if (!is_simple(x)) gr <- simplify(x)
cl <- clusters(x)
graph_splitting <- function(k, x, cl){
induced.subgraph(x, cl$membership == k)
}
components <- sapply(1:max(cl$membership), graph_splitting, x = x, cl = cl, simplify = FALSE)
}
summary.list <- lapply(components, function(x) summary(x))
}
#' @title Summarize centrality measure calculation results
#' @description This function computes the minimum, first quartile, median,
#' mean, third quartile, and maximum values of the computed centrality measures.
#' @param x Centrality measure calculation results, typically a numeric vector.
#' @return A list summarizing the computed centrality measures. The list includes the following elements:
#' - "minimum": The minimum value of the centrality measures.
#' - "first_quartile": The first quartile (25th percentile) value of the centrality measures.
#' - "median": The median (50th percentile) value of the centrality measures.
#' - "mean": The mean value of the centrality measures.
#' - "third_quartile": The third quartile (75th percentile) value of the centrality measures.
#' - "maximum": The maximum value of the centrality measures.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom stats na.omit
summary_calculate_centralities <- function(x){
x <- x[!sapply(x, is.null)]
x <- na.omit(x)
summary_list <- lapply(x, function(x) summary(x))
return(summary_list)
}
#' @title Summarize t-Distributed Stochastic Neighbor Embedding (t-SNE) on centrality measures
#'
#' @description This function summarizes the t-SNE analysis results on centrality measures.
#' @param x A list containing the computed centrality values.
#' @param dims An integer specifying the number of output dimensions (default = 2).
#' @param perplexity A numeric value representing a flexible measure of the efficient number of neighbors.
#' The performance of t-SNE is fairly robust to changes in perplexity, and typical values are between 5 and 50 (default = 5).
#' @param scale A logical value indicating whether the centrality values should be scaled or not (default = TRUE).
#' @seealso \code{\link[Rtsne]{Rtsne}}
#' @return A list containing the following elements:
#' - "Y": A matrix containing the new representations for the objects after t-SNE.
#' - "costs": The cost for every object after the final iteration of t-SNE.
#' The "costs" provide information about the optimization process.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @export
#' @importFrom Rtsne Rtsne
#' @importFrom stats na.omit
summary_tsne_centralities<-function( x , dims = 2, perplexity = 5, scale = TRUE){
x <- x[!sapply(x, is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
if (scale%in%TRUE){
x <- scale(x, center = TRUE, scale = scale)
x <- x[!duplicated(x), ]
tsne.Y <- Rtsne(t(x), dims = dims, perplexity = perplexity, check_duplicates = FALSE)$Y
rownames(tsne.Y)<-colnames(x)
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity ,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
tsne_cost <- sort(cost)
res_tsne <- list(tsne.Y = tsne.Y, tsne_cost = tsne_cost)
return(res_tsne)
}
else{
x <- x[!duplicated(x), ]
tsne.Y <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$Y
rownames(tsne.Y) <- colnames(x)
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
tsne_cost <- sort(cost)
res_tsne <- list(tsne.Y = tsne.Y,tsne_cost = tsne_cost)
return(res_tsne)
}
}
#' @title t-Distributed Stochastic Neighbor Embedding (t-SNE) on centrality measures
#'
#' @description This function applies t-SNE, a dimensionality reduction algorithm, to centrality measures.
#' @param x A list containing the computed centrality values.
#' @param dims An integer specifying the number of output dimensions (default = 2).
#' @param perplexity A numeric value representing a flexible measure of the efficient number of neighbors.
#' The performance of t-SNE is fairly robust to changes in perplexity, and typical values are between 5 and 50 (default = 5).
#' @param scale A logical value indicating whether the centrality values should be scaled or not (default = TRUE).
#' @details t-SNE is a non-linear dimensionality reduction algorithm used for exploring high-dimensional data. It maps multi-dimensional centrality measure data to a lower-dimensional space suitable for analysis and visualization.
#' @seealso \code{\link[Rtsne]{Rtsne}}
#' @return A cost plot of t-SNE results, which displays centralities in order of their corresponding costs.
#' The cost plot provides information about the optimization process and the quality of the embedding.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' van der Maaten, L. (2014). Accelerating t-SNE using Tree-Based Algorithms. Journal of Machine Learning Research, 15, 3221–3245.
#' Van Der Maaten, L. J. P., & Hinton, G. E. (2008). Visualizing High-Dimensional Data Using t-SNE. Journal of Machine Learning Research, 9, 2579–2605.
#' @export
#' @importFrom Rtsne Rtsne
#' @importFrom ggplot2 geom_bar
#' @importFrom ggplot2 guides
#' @importFrom ggplot2 xlab
#' @importFrom ggplot2 ylab
#' @importFrom ggplot2 ggtitle
#' @importFrom ggplot2 theme
#' @importFrom stats na.omit
#' @importFrom stats reorder
#' @return A cost plot of t-SNE results, which displays centralities in order of their corresponding costs.
#' The cost plot is a ggplot object that represents the optimization process and the quality of the embedding.
#' The x-axis represents the iterations of the t-SNE algorithm, and the y-axis represents the cost associated with each iteration.
#' The cost measures the discrepancy between the original high-dimensional space and the low-dimensional embedding.
#' By examining the cost plot, you can assess the convergence and stability of the t-SNE algorithm and evaluate the quality of the embedding.
tsne_centralities <- function( x , dims = 2, perplexity = 5, scale = TRUE){
x <- x[!sapply(x,is.null)]
x <- as.data.frame(x)
x <- na.omit(x)
if (scale%in%TRUE){
x <- scale(x, center = TRUE, scale = scale)
x <- x[!duplicated(x), ]
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
df <- data.frame(sort(cost))
ggplot(data=df, aes(x=reorder(rownames(df), -cost), y=cost, fill=rownames(df))) +
geom_bar(colour="black", fill="turquoise", width=.8, stat="identity") +
guides(fill=FALSE) +
xlab("Centrality measures") + ylab("tsne costs") +
ggtitle("tsne cost results")+ theme(axis.text.x=element_text(angle=45, vjust=0.5),plot.title = element_text(hjust = 0.5)
,panel.border = element_rect(colour = "black", fill=NA, size=1.5))
}
else{
x <- x[!duplicated(x), ]
cost <- Rtsne(t(x), dims = dims, perplexity = perplexity,check_duplicates = FALSE)$cost
names(cost) <- colnames(x)
df <- data.frame(sort(cost))
ggplot(data=df, aes(x=reorder(rownames(df), -cost), y=cost, fill=rownames(df))) +
geom_bar(colour="black", fill="turquoise", width=.8, stat="identity") +
guides(fill=FALSE) +
xlab("Centrality measures") + ylab("tsne costs") +
ggtitle("tsne cost results")+ theme(axis.text.x=element_text(angle=45, vjust=0.5),
plot.title = element_text(hjust = 0.5)
,panel.border = element_rect(colour = "black", fill=NA, size=1.5))
}
}
#' @title Dangalchev Closeness Centrality
#'
#' @description This function computes the Dangalchev Closeness Centrality for nodes in a network.
#' The Dangalchev Closeness Centrality measures closeness by removing nodes and edges, allowing for easier evaluation and handling of unconnected graphs.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#' @param mode A character value indicating whether the shortest paths "in" or "out" of the nodes in directed graphs should be considered. For undirected graphs, use "all".
#' @param weights A numeric vector indicating the weights of the edges.
#'
#' @seealso \code{\link[centiserve]{closeness.residual}}
#' @return A numeric vector including the centrality values for each node.
#' The centrality values represent the Dangalchev Closeness Centrality measure for each node in the network.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Dangalchev, C. (2006). Residual closeness in networks. Physica A: Statistical Mechanics and its Applications, 365, 556-564. DOI: 10.1016/j.physa.2005.12.020
#'
#' @examples
#'
#' data(zachary)
#'
#' dangalchev_closeness_centrality(zachary)
#'
#' @export
#' @importFrom igraph distances
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
#' @importFrom intergraph asIgraph
dangalchev_closeness_centrality<-function (x, vids = V(x), mode = c("all", "out", "in"), weights = NULL){
if (!("igraph" %in% class(x) || "network" %in% class(x))) stop("The input is not an igraph or a
network object")
if (is_igraph(x)){
distMat<-1/2^distances(x)
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
distMat<-1/2^distances(x)
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
return(res)
}
#' @title Local Bridging Centrality
#'
#' @description This function computes the Local Bridging Centrality for nodes in a network. The Local Bridging Centrality classifies nodes based on their structural links among the dense components.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#'
#' @seealso \code{\link[igraph]{betweenness}}
#' @return A numeric vector including the centrality values for each node.
#' The centrality values represent the Local Bridging Centrality measure for each node in the network.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Macker, J.P. (2016). An improved local bridging centrality model for distributed network analytics. In Military Communications Conference, MILCOM 2016-2016 IEEE (pp. 600-605). IEEE. DOI: 10.1109/MILCOM.2016.7795393
#'
#' @examples
#'
#' data(zachary)
#'
#' local_bridging_centrality(zachary)
#'
#' @export
#' @importFrom igraph degree
#' @importFrom igraph neighbors
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
#' @importFrom intergraph asIgraph
local_bridging_centrality<-function (x, vids = V(x)){
if (!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a
network object")
f <- function(v){
(1/degree(x,v))/sum(1/degree(x,neighbors(x,v,'all')))}
if (is_igraph(x)){
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
return(results)
}
#' @title Wiener Index Centrality
#'
#' @description This function computes the Wiener Index Centrality for nodes in a network. The Wiener index is a measure of centrality that computes the sum of all shortest paths between a node and all other related nodes in the graph. In essence, it is similar to closeness centrality, but without taking the reciprocal, resulting in a measure with the opposite meaning.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#' @param mode A character value indicating whether the shortest paths "in" or "out" of the nodes in directed graphs should be considered. For undirected graphs, "all" is used.
#' @param weights Numeric vector indicating weights of the edges.
#'
#' @return A numeric vector including the centrality values for each node.
#' The centrality values represent the Wiener Index Centrality measure for each node in the network.
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#' @references
#' Scardoni, G. and Carlo Laudanna, C.B.M.C. (2011). Network centralities for Cytoscape. University of Verona.
#'
#' @examples
#'
#' data(zachary)
#'
#' wiener_index_centrality(zachary)
#'
#' @export
#' @importFrom igraph shortest.paths
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
#' @importFrom intergraph asIgraph
wiener_index_centrality<-function (x, vids = V(x), mode = c("all", "out", "in"), weights = NULL){
if (!("igraph" %in% class(x)||"network" %in% class(x))) stop("The input is not an igraph or a
network object")
if (is_igraph(x)){
distMat<- shortest.paths(x , mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
distMat<- shortest.paths(x , mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
return(res=res)
}
#' @title Harmonic Centrality
#'
#' @description This function computes the Harmonic Centrality for nodes in a network. The harmonic centrality metric is defined as the denormalized reciprocal of the harmonic mean of all distances.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#' @param mode A character value, indicating the type of degree to consider ("out" for out-degree, "in" for in-degree, "total" for the sum of the two). For undirected graphs, this argument is ignored. The default value is "total".
#' @param weights Numeric vector indicating weights of the edges.
#'
#' @return A numeric vector of centrality values for each node. The length of the vector is equal to the number of nodes in the network.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' BOLDI, P. & VIGNA, S. 2014. Axioms for centrality. Internet Mathematics, 00-00.
#'
#' MARCHIORI, M. & LATORA, V. 2000. Harmony in the small-world. Physica A: Statistical Mechanics and its Applications, 285, 539-546.
#'
#' OPSAHL, T., AGNEESSENS, F. & SKVORETZ, J. 2010. Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32, 245-251.
#'
#' OPSAHL, T. 2010. Closeness centrality in networks with disconnected components (http://toreopsahl.com/2010/03/20/closeness-centrality-in-networks-with-disconnected-components/)
#'
#' @examples
#'
#' data(zachary)
#'
#' harmonic_centrality(zachary)
#'
#' @export
#' @importFrom igraph distances
#' @importFrom igraph vcount
#' @importFrom intergraph asIgraph
#' @importFrom igraph degree
#' @importFrom igraph neighbors
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
harmonic_centrality <- function (x, vids = V(x), mode = c("all", "out", "in"), weights = NULL){
if (!("igraph" %in% class(x) || "network" %in% class(x))) stop("The input is not an igraph or a
network object")
if (is_igraph(x)){
distMat<-1/distances(x, mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
distMat<-1/distances(x, mode = mode[1], weights = weights)
diag(distMat)<-0
res <- rowSums(distMat) - diag(distMat)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(res) <- V(x)$name[vids]
}
}
return(res)
}
#' @title Group Centrality
#'
#' @description This function computes the Group Centrality for nodes in a network. Group centrality considers a consistent ranking of each node to be calculated, taking into account the diverse possible synergies among possible groups of vertices.
#'
#' @param x An igraph or a network object.
#' @param vids Nodes to be considered in the calculation.
#'
#' @return A numeric vector of centrality values for each node. The length of the vector is equal to the number of nodes in the network.
#'
#' @author Minoo Ashtiani, Mehdi Mirzaie, Mohieddin Jafari
#'
#' @references
#' Michalak, T.P., Aadithya, K.V., Szczepanski, P.L., Ravindran, B. and Jennings, N.R., 2013. Efficient computation of the Shapley value for game-theoretic network centrality. Journal of Artificial Intelligence Research, 46, pp.607-650.
#'
#' https://www.civilica.com/Paper-IBIS07-IBIS07_127.html
#'
#' @examples
#'
#' data(zachary)
#'
#' group_centrality(zachary)
#'
#' @export
#' @importFrom igraph neighbors
#' @importFrom igraph degree
#' @importFrom igraph V
#' @importFrom igraph is_igraph
#' @importFrom igraph is_named
#' @importFrom igraph getIgraphOpt
group_centrality<-function (x, vids = V(x)){
if (!("igraph" %in% class(x)|| "network" %in% class(x))) stop("The input is not an igraph or a
network object")
f <- function(v){ 1/(1 + degree(x,v)) + sum(1/(1+degree(x,neighbors(x,v,'all'))))}
if (is_igraph(x)){
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
else{
x<-asIgraph(x)
results <- sapply(V(x),f)
if (getIgraphOpt("add.vertex.names") && is_named(x)) {
names(results) <- V(x)$name[vids]
}
}
return(results)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.