R/initClusters.details.R

In LearnClust: Learning Hierarchical Clustering Algorithms

#' @title To explain how to initialize clusters for the divisive algorithm.
#' @description To explain how to initialize clusters for the divisive algorithm.
#' @param initList is a clusters list. It will contain clusters with one element.
#' @details This function will explain how to calculate every cluster that can be created by joining initial clusters with each other. It creates clusters
#' from length = 1 until a cluster with every element is created.
#' @details These clusters will be used to find the most different clusters that we can create by dividing the initial cluster.
#' @author Roberto Alcántara \email{roberto.alcantara@@edu.uah.es}
#' @author Juan José Cuadrado \email{jjcg@@uah.es}
#' @author Universidad de Alcalá de Henares
#' @return A cluster list. Explanation.
#' @examples
#'
#' data <- c(1:8)
#'
#' matrix <- matrix(data, ncol=2)
#'
#' listData <- toListDivisive(data)
#'
#' listMatrix <- toListDivisive(matrix)
#'
#' initClusters.details(listData)
#'
#' initClusters.details(listMatrix)
#'
#' @export

initClusters.details <- function(initList){
  message("\n 'initClusters' method initializes the clusters used in the divisive algorithm. \n\n")
  message("\n To know which are the most different clusters, we need to know the distance between \n every possible clusters that could be created with the initial elements. \n")
  message("\n This step is the most computationally complex, so it will make the algorithm to get the \n solution with delay, or even, not to find a solution because of the computers capacities.\n\n")

  clusters <- initList
  unitClusters <- initList
  aux <- initList
  goal <- initList[[length(initList)]]
  res <- c()
  auxAux <- c()

  while(nrow(clusters[[length(clusters)]]) != length(unitClusters)){

    for(i in seq_len(length(aux))){
      lastElement <- aux[[i]]
      lastElement <- (lastElement[nrow(lastElement),])
      lastElement <- matrix(lastElement,ncol=2)

      clusterIndex <- getClusterIndex(unitClusters,lastElement)
      for (j in 1:length(unitClusters)){
        cluster1 <- aux[[i]]
        cluster1 <- matrix(cluster1,ncol=2)
        cluster2 <- unitClusters[[j]]
        if (j > clusterIndex){
          newCluster <- c()
          for (k in (1:nrow(cluster1))){
            newCluster <- c(newCluster,cluster1[k,])
          }
          for (k in (1:nrow(cluster2))){
            newCluster <- c(newCluster,cluster2[k,])
          }
          newCluster <- matrix(newCluster, ncol=2, byrow=TRUE)
          res[[length(res) + 1]] <- newCluster
          data <- newCluster[nrow(newCluster),]
          asMatrix <- matrix(data,ncol=2)
          if(!equalCluster(asMatrix,goal)){
            auxAux[[length(auxAux) + 1]] <- newCluster
          }
        }
      }
    }
    aux <- auxAux
    auxAux <- c()
    clusters <- c(clusters, res)
    res <- c()
  }
  message("\n The clusters created using \n")
  print(unitClusters)
  message("\n are: \n")
  print(clusters)

  clusters
}