In CRC1266-A2/moin: Modelling Interactions

library(moin)


testdata <- sf::st_read("../vignettes/data/OldenburgerGraben/Placenames.shp") 

Functions modelgraph and modelgraph2 both aim at generating graphs for interaction models.

Function modelgraph requires a two step process: 1) producing the graph using package spatgraphs with function modelgraph, 2) transforming the spatgraph object into a tidygraph object ("tbl_graph") with function spatgraph_transform(). So for a result both functions need to be exectuted.

Function modelgraph2 directly produces the graph for interaction models without requiring the execution of another function. The resulting object is a list with the tidygraph, the adjacency matrix, the distance matrix and a truncated distance matrix (see below).

# 1. model graphs with function modelgraph


# MDM - Maximum Distance Modell
mdm <- modelgraph(testdata,
                  "mdm",
                  4000)

spatgraph_transform(x=mdm,output="tidygraph")

plot(spatgraph_transform(x=mdm,output="tidygraph"),
     vertex.size=10,
     vertex.label.cex=0.5,
     layout=sf::st_coordinates(testdata))

# PPA - Proximal Point Analysis
ppa <- modelgraph(testdata,
                  "ppa",
                  4)

spatgraph_transform(x=ppa,output="tidygraph")

plot(spatgraph_transform(x=ppa,output="tidygraph"),
     vertex.size=10,
     vertex.label.cex=0.5,
     layout=sf::st_coordinates(testdata))

Now compare function modelgraph2. Note also that the interpretation of argument par differs according to the choice of argument method: for an MDM-graph, par resembles the threshold value "D", for a PPA-Graph, par resembles the rank of the neighbours to be consideren (k-th neighbour).

# 2. model graphs with function modelgraph

mdm2 <- modelgraph2(testdata, method = "mdm", par=4000)

str(mdm2)

ppa2 <- modelgraph2(testdata, method = "ppa", par=4)

plot(ppa2$graph,
     vertex.size=3,
     vertex.label.cex=0.25,
     layout=sf::st_coordinates(ppa2$input.data))

Note: function modelgraph2 by default produces a full graph for later use with Intervening Opportunity Models.

iom <- modelgraph2(testdata)

names(iom) # Don't plot the full graph ;-)

CRC1266-A2/moin documentation built on May 7, 2019, 8:56 p.m.