throw away old code, prepare tests
[epclust.git] / epclust / R / clustering.R
CommitLineData
7b13d0c2 1# Cluster one full task (nb_curves / ntasks series)
e205f218 2clusteringTask = function(indices,getSeries,getSeriesForSynchrones,synchrones_file,
3eef8d3d 3 getCoefs,K1,K2,nb_series_per_chunk,ncores,to_file,ftype)
5c652979 4{
0e2dce80 5 cl = parallel::makeCluster(ncores)
7b13d0c2
BA
6 repeat
7 {
e205f218 8 nb_workers = max( 1, round( length(indices) / nb_series_per_chunk ) )
48108c39 9 indices_workers = lapply(seq_len(nb_workers), function(i) {
7b13d0c2 10 upper_bound = ifelse( i<nb_workers,
e205f218
BA
11 min(nb_series_per_chunk*i,length(indices)), length(indices) )
12 indices[(nb_series_per_chunk*(i-1)+1):upper_bound]
48108c39 13 })
e205f218
BA
14 indices = unlist( parallel::parLapply(cl, indices_workers, function(inds)
15 computeClusters1(inds, getCoefs, K1)) )
0e2dce80 16 if (length(indices_clust) == K1)
7b13d0c2
BA
17 break
18 }
e205f218
BA
19 parallel::stopCluster(cl)
20 if (K2 == 0)
21 return (indices)
3eef8d3d
BA
22 computeClusters2(indices, K2, getSeries, getSeriesForSynchrones, to_file,
23 nb_series_per_chunk,ftype)
e205f218 24 vector("integer",0)
5c652979
BA
25}
26
0e2dce80
BA
27# Apply the clustering algorithm (PAM) on a coeffs or distances matrix
28computeClusters1 = function(indices, getCoefs, K1)
e205f218
BA
29{
30 coefs = getCoefs(indices)
31 indices[ cluster::pam(coefs, K1, diss=FALSE)$id.med ]
32}
0e2dce80 33
7b13d0c2 34# Cluster a chunk of series inside one task (~max nb_series_per_chunk)
3eef8d3d
BA
35computeClusters2 = function(indices, K2, getSeries, getSeriesForSynchrones, to_file,
36 nb_series_per_chunk, ftype)
5c652979 37{
3eef8d3d 38 curves = computeSynchrones(indices, getSeries, getSeriesForSynchrones, nb_series_per_chunk)
e205f218
BA
39 dists = computeWerDists(curves)
40 medoids = cluster::pam(dists, K2, diss=TRUE)$medoids
41 if (to_file)
5c652979 42 {
3eef8d3d 43 serialize(medoids, synchrones_file, ftype, nb_series_per_chunk)
e205f218 44 return (NULL)
5c652979 45 }
e205f218 46 medoids
5c652979
BA
47}
48
7b13d0c2 49# Compute the synchrones curves (sum of clusters elements) from a clustering result
3eef8d3d 50computeSynchrones = function(indices, getSeries, getSeriesForSynchrones, nb_series_per_chunk)
e205f218
BA
51{
52 #les getSeries(indices) sont les medoides --> init vect nul pour chacun, puis incr avec les
53 #courbes (getSeriesForSynchrones) les plus proches... --> au sens de la norme L2 ?
3eef8d3d
BA
54 medoids = getSeries(indices)
55 K = nrow(medoids)
56 synchrones = matrix(0, nrow=K, ncol=ncol(medoids))
57 counts = rep(0,K)
58 index = 1
59 repeat
60 {
61 series = getSeriesForSynchrones((index-1)+seq_len(nb_series_per_chunk))
62 if (is.null(series))
63 break
64 #get medoids indices for this chunk of series
65 index = which.min( rowSums( sweep(medoids, 2, series[i,], '-')^2 ) )
66 synchrones[index,] = synchrones[index,] + series[i,]
67 counts[index] = counts[index] + 1
68 }
69 #NOTE: odds for some clusters to be empty? (when series already come from stage 2)
70 synchrones = sweep(synchrones, 1, counts, '/')
e205f218 71}
1c6f223e 72
e205f218 73# Compute the WER distance between the synchrones curves (in rows)
7b13d0c2 74computeWerDist = function(curves)
d03c0621 75{
5c652979
BA
76 if (!require("Rwave", quietly=TRUE))
77 stop("Unable to load Rwave library")
7b13d0c2
BA
78 n <- nrow(curves)
79 delta <- ncol(curves)
db6fc17d 80 #TODO: automatic tune of all these parameters ? (for other users)
d03c0621 81 nvoice <- 4
7b13d0c2 82 # noctave = 2^13 = 8192 half hours ~ 180 days ; ~log2(ncol(curves))
d7d55bc1
BA
83 noctave = 13
84 # 4 here represent 2^5 = 32 half-hours ~ 1 day
db6fc17d
BA
85 #NOTE: default scalevector == 2^(0:(noctave * nvoice) / nvoice) * s0 (?)
86 scalevector <- 2^(4:(noctave * nvoice) / nvoice) * 2
87 #condition: ( log2(s0*w0/(2*pi)) - 1 ) * nvoice + 1.5 >= 1
88 s0=2
89 w0=2*pi
90 scaled=FALSE
91 s0log = as.integer( (log2( s0*w0/(2*pi) ) - 1) * nvoice + 1.5 )
92 totnoct = noctave + as.integer(s0log/nvoice) + 1
93
94 # (normalized) observations node with CWT
95 Xcwt4 <- lapply(seq_len(n), function(i) {
e205f218 96 ts <- scale(ts(curves[i,]), center=TRUE, scale=scaled)
db6fc17d
BA
97 totts.cwt = Rwave::cwt(ts,totnoct,nvoice,w0,plot=0)
98 ts.cwt = totts.cwt[,s0log:(s0log+noctave*nvoice)]
99 #Normalization
100 sqs <- sqrt(2^(0:(noctave*nvoice)/nvoice)*s0)
101 sqres <- sweep(ts.cwt,MARGIN=2,sqs,'*')
102 sqres / max(Mod(sqres))
103 })
3ccd1e39 104
db6fc17d
BA
105 Xwer_dist <- matrix(0., n, n)
106 fcoefs = rep(1/3, 3) #moving average on 3 values (TODO: very slow! correct?!)
107 for (i in 1:(n-1))
1c6f223e 108 {
db6fc17d 109 for (j in (i+1):n)
d03c0621 110 {
0e2dce80 111 #TODO: later, compute CWT here (because not enough storage space for 200k series)
db6fc17d
BA
112 # 'circular=TRUE' is wrong, should just take values on the sides; to rewrite in C
113 num <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE)
114 WX <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[i]])), fcoefs, circular=TRUE)
115 WY <- filter(Mod(Xcwt4[[j]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE)
116 wer2 <- sum(colSums(num)^2) / sum( sum(colSums(WX) * colSums(WY)) )
117 Xwer_dist[i,j] <- sqrt(delta * ncol(Xcwt4[[1]]) * (1 - wer2))
118 Xwer_dist[j,i] <- Xwer_dist[i,j]
d03c0621 119 }
1c6f223e 120 }
d03c0621 121 diag(Xwer_dist) <- numeric(n)
c6556868 122 Xwer_dist
1c6f223e 123}