epclust/R/clustering.R

   1 #' Two-stage clustering, within one task (see \code{claws()})
   2 #'
   3 #' \code{clusteringTask1()} runs one full stage-1 task, which consists in iterated
   4 #' clustering on nb_curves / ntasks energy contributions, computed through
   5 #' discrete wavelets coefficients.
   6 #' \code{clusteringTask2()} runs a full stage-2 task, which consists in WER distances
   7 #' computations between medoids (indices) output from stage 1, before applying
   8 #' the second clustering algorithm on the distances matrix.
   9 #'
  10 #' @param getContribs Function to retrieve contributions from initial series indices:
  11 #'   \code{getContribs(indices)} outputs a contributions matrix, in columns
  12 #' @inheritParams claws
  13 #' @inheritParams computeSynchrones
  14 #' @inheritParams computeWerDists
  15 #'
  16 #' @return The indices of the computed (resp. K1 and K2) medoids.
  17 #'
  18 #' @name clustering
  19 #' @rdname clustering
  20 #' @aliases clusteringTask1 clusteringTask2
  21 NULL
  22
  23 #' @rdname clustering
  24 #' @export
  25 clusteringTask1 <- function(indices, getContribs, K1, algoClust1, nb_items_clust,
  26     ncores_clust=3, verbose=FALSE, parll=TRUE)
  27 {
  28     if (verbose)
  29         cat(paste("*** Clustering task 1 on ",length(indices)," series [start]\n", sep=""))
  30
  31     if (length(indices) <= K1)
  32         return (indices)
  33
  34     if (parll)
  35     {
  36         # outfile=="" to see stderr/stdout on terminal
  37         cl <-
  38             if (verbose)
  39                 parallel::makeCluster(ncores_clust, outfile = "")
  40             else
  41                 parallel::makeCluster(ncores_clust)
  42         parallel::clusterExport(cl, c("getContribs","K1","verbose"), envir=environment())
  43     }
  44     # Iterate clustering algorithm 1 until K1 medoids are found
  45     while (length(indices) > K1)
  46     {
  47         # Balance tasks by splitting the indices set - as evenly as possible
  48         indices_workers <- .splitIndices(indices, nb_items_clust, min_size=K1+1)
  49         indices <-
  50             if (parll)
  51             {
  52                 unlist( parallel::parLapply(cl, indices_workers, function(inds) {
  53                     require("epclust", quietly=TRUE)
  54                     inds[ algoClust1(getContribs(inds), K1) ]
  55                 }) )
  56             }
  57             else
  58             {
  59                 unlist( lapply(indices_workers, function(inds)
  60                     inds[ algoClust1(getContribs(inds), K1) ]
  61                 ) )
  62             }
  63         if (verbose)
  64         {
  65             cat(paste("*** Clustering task 1 on ",length(indices)," medoids [iter]\n", sep=""))
  66         }
  67     }
  68     if (parll)
  69         parallel::stopCluster(cl)
  70
  71     indices #medoids
  72 }
  73
  74 #' @rdname clustering
  75 #' @export
  76 clusteringTask2 <- function(indices, getSeries, K2, algoClust2, nb_series_per_chunk,
  77     smooth_lvl, nvoice, nbytes, endian, ncores_clust=3, verbose=FALSE, parll=TRUE)
  78 {
  79     if (verbose)
  80         cat(paste("*** Clustering task 2 on ",length(indices)," medoids\n", sep=""))
  81
  82     if (length(indices) <= K2)
  83         return (indices)
  84
  85     # A) Compute the WER distances (Wavelets Extended coefficient of deteRmination)
  86     distances <- computeWerDists(indices, getSeries, nb_series_per_chunk,
  87         smooth_lvl, nvoice, nbytes, endian, ncores_clust, verbose, parll)
  88
  89     # B) Apply clustering algorithm 2 on the WER distances matrix
  90     if (verbose)
  91         cat(paste("*** algoClust2() on ",nrow(distances)," items\n", sep=""))
  92     indices[ algoClust2(distances,K2) ]
  93 }