'update'
[epclust.git] / epclust / R / clustering.R
CommitLineData
56857861
BA
1# Cluster one full task (nb_curves / ntasks series); only step 1
2clusteringTask = function(indices, getCoefs, K1, nb_series_per_chunk, ncores)
5c652979 3{
0e2dce80 4 cl = parallel::makeCluster(ncores)
8702eb86 5 parallel::clusterExport(cl, varlist=c("getCoefs","K1"), envir=environment())
7b13d0c2
BA
6 repeat
7 {
8702eb86
BA
8 nb_workers = max( 1, floor( length(indices) / nb_series_per_chunk ) )
9 indices_workers = lapply( seq_len(nb_workers), function(i)
10 indices[(nb_series_per_chunk*(i-1)+1):(nb_series_per_chunk*i)] )
11 # Spread the remaining load among the workers
12 rem = length(indices) %% nb_series_per_chunk
13 while (rem > 0)
14 {
15 index = rem%%nb_workers + 1
16 indices_workers[[index]] = c(indices_workers[[index]], tail(indices,rem))
17 rem = rem - 1
18 }
19 indices = unlist( parallel::parLapply( cl, indices_workers, function(inds) {
20 require("epclust", quietly=TRUE)
21 inds[ computeClusters1(getCoefs(inds), K1) ]
22 } ) )
56857861 23 if (length(indices) == K1)
7b13d0c2
BA
24 break
25 }
e205f218 26 parallel::stopCluster(cl)
56857861 27 indices #medoids
5c652979
BA
28}
29
0e2dce80 30# Apply the clustering algorithm (PAM) on a coeffs or distances matrix
56857861 31computeClusters1 = function(coefs, K1)
8702eb86 32 cluster::pam(coefs, K1, diss=FALSE)$id.med
0e2dce80 33
7b13d0c2 34# Cluster a chunk of series inside one task (~max nb_series_per_chunk)
56857861 35computeClusters2 = function(medoids, K2, getRefSeries, nb_series_per_chunk)
5c652979 36{
56857861 37 synchrones = computeSynchrones(medoids, getRefSeries, nb_series_per_chunk)
8702eb86 38 medoids[ cluster::pam(computeWerDists(synchrones), K2, diss=TRUE)$medoids , ]
5c652979
BA
39}
40
7b13d0c2 41# Compute the synchrones curves (sum of clusters elements) from a clustering result
56857861 42computeSynchrones = function(medoids, getRefSeries, nb_series_per_chunk)
e205f218 43{
3eef8d3d
BA
44 K = nrow(medoids)
45 synchrones = matrix(0, nrow=K, ncol=ncol(medoids))
46 counts = rep(0,K)
47 index = 1
48 repeat
49 {
56857861
BA
50 range = (index-1) + seq_len(nb_series_per_chunk)
51 ref_series = getRefSeries(range)
52 if (is.null(ref_series))
3eef8d3d
BA
53 break
54 #get medoids indices for this chunk of series
56857861
BA
55 for (i in seq_len(nrow(ref_series)))
56 {
8702eb86
BA
57 j = which.min( rowSums( sweep(medoids, 2, ref_series[i,], '-')^2 ) )
58 synchrones[j,] = synchrones[j,] + ref_series[i,]
56857861
BA
59 counts[j] = counts[j] + 1
60 }
61 index = index + nb_series_per_chunk
3eef8d3d
BA
62 }
63 #NOTE: odds for some clusters to be empty? (when series already come from stage 2)
8702eb86
BA
64 # ...maybe; but let's hope resulting K1' be still quite bigger than K2
65 synchrones = sweep(synchrones, 1, counts, '/')
66 synchrones[ sapply(seq_len(K), function(i) all(!is.nan(synchrones[i,]))) , ]
e205f218 67}
1c6f223e 68
e205f218 69# Compute the WER distance between the synchrones curves (in rows)
8702eb86 70computeWerDists = function(curves)
d03c0621 71{
5c652979
BA
72 if (!require("Rwave", quietly=TRUE))
73 stop("Unable to load Rwave library")
7b13d0c2
BA
74 n <- nrow(curves)
75 delta <- ncol(curves)
db6fc17d 76 #TODO: automatic tune of all these parameters ? (for other users)
d03c0621 77 nvoice <- 4
7b13d0c2 78 # noctave = 2^13 = 8192 half hours ~ 180 days ; ~log2(ncol(curves))
d7d55bc1
BA
79 noctave = 13
80 # 4 here represent 2^5 = 32 half-hours ~ 1 day
db6fc17d
BA
81 #NOTE: default scalevector == 2^(0:(noctave * nvoice) / nvoice) * s0 (?)
82 scalevector <- 2^(4:(noctave * nvoice) / nvoice) * 2
83 #condition: ( log2(s0*w0/(2*pi)) - 1 ) * nvoice + 1.5 >= 1
84 s0=2
85 w0=2*pi
86 scaled=FALSE
87 s0log = as.integer( (log2( s0*w0/(2*pi) ) - 1) * nvoice + 1.5 )
88 totnoct = noctave + as.integer(s0log/nvoice) + 1
89
90 # (normalized) observations node with CWT
91 Xcwt4 <- lapply(seq_len(n), function(i) {
e205f218 92 ts <- scale(ts(curves[i,]), center=TRUE, scale=scaled)
db6fc17d
BA
93 totts.cwt = Rwave::cwt(ts,totnoct,nvoice,w0,plot=0)
94 ts.cwt = totts.cwt[,s0log:(s0log+noctave*nvoice)]
95 #Normalization
96 sqs <- sqrt(2^(0:(noctave*nvoice)/nvoice)*s0)
97 sqres <- sweep(ts.cwt,MARGIN=2,sqs,'*')
98 sqres / max(Mod(sqres))
99 })
3ccd1e39 100
db6fc17d
BA
101 Xwer_dist <- matrix(0., n, n)
102 fcoefs = rep(1/3, 3) #moving average on 3 values (TODO: very slow! correct?!)
103 for (i in 1:(n-1))
1c6f223e 104 {
db6fc17d 105 for (j in (i+1):n)
d03c0621 106 {
0e2dce80 107 #TODO: later, compute CWT here (because not enough storage space for 200k series)
db6fc17d
BA
108 # 'circular=TRUE' is wrong, should just take values on the sides; to rewrite in C
109 num <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE)
110 WX <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[i]])), fcoefs, circular=TRUE)
111 WY <- filter(Mod(Xcwt4[[j]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE)
112 wer2 <- sum(colSums(num)^2) / sum( sum(colSums(WX) * colSums(WY)) )
113 Xwer_dist[i,j] <- sqrt(delta * ncol(Xcwt4[[1]]) * (1 - wer2))
114 Xwer_dist[j,i] <- Xwer_dist[i,j]
d03c0621 115 }
1c6f223e 116 }
d03c0621 117 diag(Xwer_dist) <- numeric(n)
c6556868 118 Xwer_dist
1c6f223e 119}