complete first draft of package

[epclust.git] / epclust / R / clustering.R

diff --git a/epclust/R/clustering.R b/epclust/R/clustering.R
index c8bad66..87a5f91 100644 (file)
--- a/epclust/R/clustering.R
+++ b/epclust/R/clustering.R
@@ -1,6 +1,6 @@
 # Cluster one full task (nb_curves / ntasks series)
 clusteringTask = function(indices,getSeries,getSeriesForSynchrones,synchrones_file,
-       getCoefs,K1,K2,nb_series_per_chunk,ncores,to_file)
+       getCoefs,K1,K2,nb_series_per_chunk,ncores,to_file,ftype)
 {
        cl = parallel::makeCluster(ncores)
        repeat
@@ -19,7 +19,8 @@ clusteringTask = function(indices,getSeries,getSeriesForSynchrones,synchrones_fi
        parallel::stopCluster(cl)
        if (K2 == 0)
                return (indices)
-       computeClusters2(indices, K2, getSeries, getSeriesForSynchrones, to_file)
+       computeClusters2(indices, K2, getSeries, getSeriesForSynchrones, to_file,
+                                                                        nb_series_per_chunk,ftype)
        vector("integer",0)
 }
 
@@ -31,27 +32,42 @@ computeClusters1 = function(indices, getCoefs, K1)
 }
 
 # Cluster a chunk of series inside one task (~max nb_series_per_chunk)
-computeClusters2 = function(indices, K2, getSeries, getSeriesForSynchrones, to_file)
+computeClusters2 = function(indices, K2, getSeries, getSeriesForSynchrones, to_file,
+                                                                                                               nb_series_per_chunk, ftype)
 {
-       curves = computeSynchrones(indices, getSeries, getSeriesForSynchrones)
+       curves = computeSynchrones(indices, getSeries, getSeriesForSynchrones, nb_series_per_chunk)
        dists = computeWerDists(curves)
        medoids = cluster::pam(dists, K2, diss=TRUE)$medoids
        if (to_file)
        {
-               serialize(medoids, synchrones_file)
+               serialize(medoids, synchrones_file, ftype, nb_series_per_chunk)
                return (NULL)
        }
        medoids
 }
 
 # Compute the synchrones curves (sum of clusters elements) from a clustering result
-computeSynchrones = function(indices, getSeries, getSeriesForSynchrones)
+computeSynchrones = function(indices, getSeries, getSeriesForSynchrones, nb_series_per_chunk)
 {
        #les getSeries(indices) sont les medoides --> init vect nul pour chacun, puis incr avec les
        #courbes (getSeriesForSynchrones) les plus proches... --> au sens de la norme L2 ?
-       series = getSeries(indices)
-       #...........
-       #sapply(seq_along(inds), colMeans(getSeries(inds[[i]]$indices,inds[[i]]$ids)))
+       medoids = getSeries(indices)
+       K = nrow(medoids)
+       synchrones = matrix(0, nrow=K, ncol=ncol(medoids))
+       counts = rep(0,K)
+       index = 1
+       repeat
+       {
+               series = getSeriesForSynchrones((index-1)+seq_len(nb_series_per_chunk))
+               if (is.null(series))
+                       break
+               #get medoids indices for this chunk of series
+               index = which.min( rowSums( sweep(medoids, 2, series[i,], '-')^2 ) )
+               synchrones[index,] = synchrones[index,] + series[i,]
+               counts[index] = counts[index] + 1
+       }
+       #NOTE: odds for some clusters to be empty? (when series already come from stage 2)
+       synchrones = sweep(synchrones, 1, counts, '/')
 }
 
 # Compute the WER distance between the synchrones curves (in rows)