Commit | Line | Data |
---|---|---|
7b13d0c2 BA |
1 | # Cluster one full task (nb_curves / ntasks series) |
2 | clusteringTask = function(K1, K2, WER, nb_series_per_chunk, indices_tasks, ncores_clust) | |
5c652979 | 3 | { |
7b13d0c2 BA |
4 | cl_clust = parallel::makeCluster(ncores_clust) |
5 | #parallel::clusterExport(cl=cl_clust, varlist=c("fonctions_du_package"), envir=environment()) | |
6 | indices_clust = indices_task[[i]] | |
7 | repeat | |
8 | { | |
9 | nb_workers = max( 1, round( length(indices_clust) / nb_series_per_chunk ) ) | |
10 | indices_workers = list() | |
11 | for (i in 1:nb_workers) | |
5c652979 | 12 | { |
7b13d0c2 BA |
13 | upper_bound = ifelse( i<nb_workers, |
14 | min(nb_series_per_chunk*i,length(indices_clust)), length(indices_clust) ) | |
15 | indices_workers[[i]] = indices_clust[(nb_series_per_chunk*(i-1)+1):upper_bound] | |
5c652979 | 16 | } |
7b13d0c2 BA |
17 | indices_clust = parallel::parLapply(cl, indices_workers, clusterChunk, K1, K2*(WER=="mix")) |
18 | if ((WER=="end" && length(indices_clust)==K1) || (WER=="mix" && length(indices_clust)==K2)) | |
19 | break | |
20 | } | |
21 | parallel::stopCluster(cl_clust) | |
22 | unlist(indices_clust) | |
5c652979 BA |
23 | } |
24 | ||
7b13d0c2 BA |
25 | # Cluster a chunk of series inside one task (~max nb_series_per_chunk) |
26 | clusterChunk = function(indices, K1, K2) | |
5c652979 | 27 | { |
5c652979 | 28 | coeffs = getCoeffs(indices) |
7b13d0c2 | 29 | cl = computeClusters(as.matrix(coeffs[,2:ncol(coeffs)]), K1, diss=FALSE) |
5c652979 BA |
30 | if (K2 > 0) |
31 | { | |
32 | curves = computeSynchrones(cl) | |
33 | dists = computeWerDists(curves) | |
7b13d0c2 | 34 | cl = computeClusters(dists, K2, diss=TRUE) |
5c652979 | 35 | } |
7b13d0c2 | 36 | indices[cl] |
5c652979 BA |
37 | } |
38 | ||
7b13d0c2 BA |
39 | # Apply the clustering algorithm (PAM) on a coeffs or distances matrix |
40 | computeClusters = function(md, K, diss) | |
5c652979 | 41 | { |
7b13d0c2 BA |
42 | if (!require(cluster, quietly=TRUE)) |
43 | stop("Unable to load cluster library") | |
44 | cluster::pam(md, K, diss=diss)$id.med | |
5c652979 BA |
45 | } |
46 | ||
7b13d0c2 BA |
47 | # Compute the synchrones curves (sum of clusters elements) from a clustering result |
48 | computeSynchrones = function(indices) | |
5c652979 | 49 | { |
7b13d0c2 | 50 | colSums( getData(indices) ) |
5c652979 | 51 | } |
1c6f223e | 52 | |
7b13d0c2 BA |
53 | # Compute the WER distance between the synchrones curves |
54 | computeWerDist = function(curves) | |
d03c0621 | 55 | { |
5c652979 BA |
56 | if (!require("Rwave", quietly=TRUE)) |
57 | stop("Unable to load Rwave library") | |
7b13d0c2 BA |
58 | n <- nrow(curves) |
59 | delta <- ncol(curves) | |
db6fc17d | 60 | #TODO: automatic tune of all these parameters ? (for other users) |
d03c0621 | 61 | nvoice <- 4 |
7b13d0c2 | 62 | # noctave = 2^13 = 8192 half hours ~ 180 days ; ~log2(ncol(curves)) |
d7d55bc1 BA |
63 | noctave = 13 |
64 | # 4 here represent 2^5 = 32 half-hours ~ 1 day | |
db6fc17d BA |
65 | #NOTE: default scalevector == 2^(0:(noctave * nvoice) / nvoice) * s0 (?) |
66 | scalevector <- 2^(4:(noctave * nvoice) / nvoice) * 2 | |
67 | #condition: ( log2(s0*w0/(2*pi)) - 1 ) * nvoice + 1.5 >= 1 | |
68 | s0=2 | |
69 | w0=2*pi | |
70 | scaled=FALSE | |
71 | s0log = as.integer( (log2( s0*w0/(2*pi) ) - 1) * nvoice + 1.5 ) | |
72 | totnoct = noctave + as.integer(s0log/nvoice) + 1 | |
73 | ||
74 | # (normalized) observations node with CWT | |
75 | Xcwt4 <- lapply(seq_len(n), function(i) { | |
7b13d0c2 | 76 | ts <- scale(ts(curves[i,]), center=TRUE, scale=scaled) |
db6fc17d BA |
77 | totts.cwt = Rwave::cwt(ts,totnoct,nvoice,w0,plot=0) |
78 | ts.cwt = totts.cwt[,s0log:(s0log+noctave*nvoice)] | |
79 | #Normalization | |
80 | sqs <- sqrt(2^(0:(noctave*nvoice)/nvoice)*s0) | |
81 | sqres <- sweep(ts.cwt,MARGIN=2,sqs,'*') | |
82 | sqres / max(Mod(sqres)) | |
83 | }) | |
3ccd1e39 | 84 | |
db6fc17d BA |
85 | Xwer_dist <- matrix(0., n, n) |
86 | fcoefs = rep(1/3, 3) #moving average on 3 values (TODO: very slow! correct?!) | |
87 | for (i in 1:(n-1)) | |
1c6f223e | 88 | { |
db6fc17d | 89 | for (j in (i+1):n) |
d03c0621 | 90 | { |
db6fc17d BA |
91 | #TODO: later, compute CWT here (because not enough storage space for 32M series) |
92 | # 'circular=TRUE' is wrong, should just take values on the sides; to rewrite in C | |
93 | num <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE) | |
94 | WX <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[i]])), fcoefs, circular=TRUE) | |
95 | WY <- filter(Mod(Xcwt4[[j]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE) | |
96 | wer2 <- sum(colSums(num)^2) / sum( sum(colSums(WX) * colSums(WY)) ) | |
97 | Xwer_dist[i,j] <- sqrt(delta * ncol(Xcwt4[[1]]) * (1 - wer2)) | |
98 | Xwer_dist[j,i] <- Xwer_dist[i,j] | |
d03c0621 | 99 | } |
1c6f223e | 100 | } |
d03c0621 | 101 | diag(Xwer_dist) <- numeric(n) |
c6556868 | 102 | Xwer_dist |
1c6f223e | 103 | } |