Commit | Line | Data |
---|---|---|
8702eb86 | 1 | #' CLAWS: CLustering with wAvelets and Wer distanceS |
7f0781b7 | 2 | #' |
eef6f6c9 BA |
3 | #' Cluster electricity power curves (or any series of similar nature) by applying a |
4 | #' two stage procedure in parallel (see details). | |
5 | #' Input series must be sampled on the same time grid, no missing values. | |
6 | #' | |
d9bb53c5 BA |
7 | #' Summary of the function execution flow: |
8 | #' \enumerate{ | |
9 | #' \item Compute and serialize all contributions, obtained through discrete wavelet | |
10 | #' decomposition (see Antoniadis & al. [2013]) | |
11 | #' \item Divide series into \code{ntasks} groups to process in parallel. In each task: | |
eef6f6c9 | 12 | #' \enumerate{ |
d9bb53c5 BA |
13 | #' \item iterate the first clustering algorithm on its aggregated outputs, |
14 | #' on inputs of size \code{nb_items_clust1} | |
15 | #' \item optionally, if WER=="mix": | |
16 | #' a) compute the K1 synchrones curves, | |
17 | #' b) compute WER distances (K1xK1 matrix) between synchrones and | |
18 | #' c) apply the second clustering algorithm | |
eef6f6c9 | 19 | #' } |
d9bb53c5 BA |
20 | #' \item Launch a final task on the aggregated outputs of all previous tasks: |
21 | #' in the case WER=="end" this task takes indices in input, otherwise | |
22 | #' (medoid) curves | |
23 | #' } | |
24 | #' \cr | |
25 | #' The main argument -- \code{getSeries} -- has a quite misleading name, since it can be | |
26 | #' either a [big.]matrix, a CSV file, a connection or a user function to retrieve | |
27 | #' series; the name was chosen because all types of arguments are converted to a function. | |
28 | #' When \code{getSeries} is given as a function, it must take a single argument, | |
29 | #' 'indices', integer vector equal to the indices of the curves to retrieve; | |
a52836b2 BA |
30 | #' see SQLite example. The nature and role of other arguments should be clear. |
31 | #' WARNING: the return value must be a matrix (in columns), or NULL if no matches. | |
d9bb53c5 BA |
32 | #' \cr |
33 | #' Note: Since we don't make assumptions on initial data, there is a possibility that | |
34 | #' even when serialized, contributions or synchrones do not fit in RAM. For example, | |
35 | #' 30e6 series of length 100,000 would lead to a +4Go contribution matrix. Therefore, | |
36 | #' it's safer to place these in (binary) files; that's what we do. | |
7f0781b7 | 37 | #' |
8702eb86 BA |
38 | #' @param getSeries Access to the (time-)series, which can be of one of the three |
39 | #' following types: | |
40 | #' \itemize{ | |
eef6f6c9 | 41 | #' \item [big.]matrix: each column contains the (time-ordered) values of one time-serie |
bf5c0844 BA |
42 | #' \item connection: any R connection object providing lines as described above |
43 | #' \item character: name of a CSV file containing series in rows (no header) | |
8702eb86 | 44 | #' \item function: a custom way to retrieve the curves; it has only one argument: |
eef6f6c9 | 45 | #' the indices of the series to be retrieved. See SQLite example |
8702eb86 | 46 | #' } |
eef6f6c9 | 47 | #' @param K1 Number of clusters to be found after stage 1 (K1 << N [number of series]) |
1c6f223e | 48 | #' @param K2 Number of clusters to be found after stage 2 (K2 << K1) |
37c82bba | 49 | #' @param nb_series_per_chunk (Maximum) number of series to retrieve in one batch |
0486fbad | 50 | #' @param algoClust1 Clustering algorithm for stage 1. A function which takes (data, K) |
2b9f5356 | 51 | #' as argument where data is a matrix in columns and K the desired number of clusters, |
9f05a4a0 BA |
52 | #' and outputs K medoids ranks. Default: PAM. In our method, this function is called |
53 | #' on iterated medoids during stage 1 | |
0486fbad | 54 | #' @param algoClust2 Clustering algorithm for stage 2. A function which takes (dists, K) |
2b9f5356 | 55 | #' as argument where dists is a matrix of distances and K the desired number of clusters, |
9f05a4a0 BA |
56 | #' and outputs K medoids ranks. Default: PAM. In our method, this function is called |
57 | #' on a matrix of K1 x K1 (WER) distances computed between synchrones | |
37c82bba BA |
58 | #' @param nb_items_clust1 (~Maximum) number of items in input of the clustering algorithm |
59 | #' for stage 1. At worst, a clustering algorithm might be called with ~2*nb_items_clust1 | |
60 | #' items; but this could only happen at the last few iterations. | |
eef6f6c9 BA |
61 | #' @param wav_filt Wavelet transform filter; see ?wavelets::wt.filter |
62 | #' @param contrib_type Type of contribution: "relative", "logit" or "absolute" (any prefix) | |
63 | #' @param WER "end" to apply stage 2 after stage 1 has fully iterated, or "mix" to apply | |
64 | #' stage 2 at the end of each task | |
a52836b2 | 65 | #' @param nvoice Number of voices within each octave for CWT computations |
4bcfdbee | 66 | #' @param random TRUE (default) for random chunks repartition |
eef6f6c9 BA |
67 | #' @param ntasks Number of tasks (parallel iterations to obtain K1 [if WER=="end"] |
68 | #' or K2 [if WER=="mix"] medoids); default: 1. | |
69 | #' Note: ntasks << N (number of series), so that N is "roughly divisible" by ntasks | |
70 | #' @param ncores_tasks Number of parallel tasks (1 to disable: sequential tasks) | |
71 | #' @param ncores_clust Number of parallel clusterings in one task (4 should be a minimum) | |
4bcfdbee | 72 | #' @param sep Separator in CSV input file (if any provided) |
8702eb86 | 73 | #' @param nbytes Number of bytes to serialize a floating-point number; 4 or 8 |
eef6f6c9 | 74 | #' @param endian Endianness for (de)serialization ("little" or "big") |
4bcfdbee | 75 | #' @param verbose Level of verbosity (0/FALSE for nothing or 1/TRUE for all; devel stage) |
492cd9e7 | 76 | #' @param parll TRUE to fully parallelize; otherwise run sequentially (debug, comparison) |
7f0781b7 | 77 | #' |
eef6f6c9 BA |
78 | #' @return A matrix of the final K2 medoids curves, in columns |
79 | #' | |
80 | #' @references Clustering functional data using Wavelets [2013]; | |
81 | #' A. Antoniadis, X. Brossat, J. Cugliari & J.-M. Poggi. | |
82 | #' Inter. J. of Wavelets, Multiresolution and Information Procesing, | |
83 | #' vol. 11, No 1, pp.1-30. doi:10.1142/S0219691313500033 | |
1c6f223e BA |
84 | #' |
85 | #' @examples | |
4efef8cc | 86 | #' \dontrun{ |
eef6f6c9 | 87 | #' # WER distances computations are too long for CRAN (for now) |
4efef8cc BA |
88 | #' |
89 | #' # Random series around cos(x,2x,3x)/sin(x,2x,3x) | |
90 | #' x = seq(0,500,0.05) | |
91 | #' L = length(x) #10001 | |
eef6f6c9 | 92 | #' ref_series = matrix( c(cos(x),cos(2*x),cos(3*x),sin(x),sin(2*x),sin(3*x)), ncol=6 ) |
4efef8cc | 93 | #' library(wmtsa) |
eef6f6c9 | 94 | #' series = do.call( cbind, lapply( 1:6, function(i) |
a52836b2 | 95 | #' do.call(cbind, wmtsa::wavBootstrap(ref_series[,i], n.realization=400)) ) ) |
4efef8cc | 96 | #' #dim(series) #c(2400,10001) |
37c82bba | 97 | #' medoids_ascii = claws(series, K1=60, K2=6, 200, verbose=TRUE) |
4efef8cc BA |
98 | #' |
99 | #' # Same example, from CSV file | |
100 | #' csv_file = "/tmp/epclust_series.csv" | |
101 | #' write.table(series, csv_file, sep=",", row.names=FALSE, col.names=FALSE) | |
37c82bba | 102 | #' medoids_csv = claws(csv_file, K1=60, K2=6, 200) |
4efef8cc BA |
103 | #' |
104 | #' # Same example, from binary file | |
eef6f6c9 BA |
105 | #' bin_file <- "/tmp/epclust_series.bin" |
106 | #' nbytes <- 8 | |
107 | #' endian <- "little" | |
108 | #' binarize(csv_file, bin_file, 500, nbytes, endian) | |
109 | #' getSeries <- function(indices) getDataInFile(indices, bin_file, nbytes, endian) | |
37c82bba | 110 | #' medoids_bin <- claws(getSeries, K1=60, K2=6, 200) |
4efef8cc BA |
111 | #' unlink(csv_file) |
112 | #' unlink(bin_file) | |
113 | #' | |
114 | #' # Same example, from SQLite database | |
115 | #' library(DBI) | |
116 | #' series_db <- dbConnect(RSQLite::SQLite(), "file::memory:") | |
117 | #' # Prepare data.frame in DB-format | |
eef6f6c9 BA |
118 | #' n <- nrow(series) |
119 | #' time_values <- data.frame( | |
4bcfdbee BA |
120 | #' id = rep(1:n,each=L), |
121 | #' time = rep( as.POSIXct(1800*(0:n),"GMT",origin="2001-01-01"), L ), | |
122 | #' value = as.double(t(series)) ) | |
4efef8cc | 123 | #' dbWriteTable(series_db, "times_values", times_values) |
4bcfdbee BA |
124 | #' # Fill associative array, map index to identifier |
125 | #' indexToID_inDB <- as.character( | |
126 | #' dbGetQuery(series_db, 'SELECT DISTINCT id FROM time_values')[,"id"] ) | |
eef6f6c9 BA |
127 | #' serie_length <- as.integer( dbGetQuery(series_db, |
128 | #' paste("SELECT COUNT * FROM time_values WHERE id == ",indexToID_inDB[1],sep="")) ) | |
129 | #' getSeries <- function(indices) { | |
130 | #' request <- "SELECT id,value FROM times_values WHERE id in (" | |
4bcfdbee | 131 | #' for (i in indices) |
eef6f6c9 BA |
132 | #' request <- paste(request, indexToID_inDB[i], ",", sep="") |
133 | #' request <- paste(request, ")", sep="") | |
134 | #' df_series <- dbGetQuery(series_db, request) | |
a52836b2 BA |
135 | #' if (length(df_series) >= 1) |
136 | #' as.matrix(df_series[,"value"], nrow=serie_length) | |
137 | #' else | |
138 | #' NULL | |
4efef8cc | 139 | #' } |
37c82bba | 140 | #' medoids_db = claws(getSeries, K1=60, K2=6, 200)) |
4bcfdbee BA |
141 | #' dbDisconnect(series_db) |
142 | #' | |
143 | #' # All computed medoids should be the same: | |
144 | #' digest::sha1(medoids_ascii) | |
145 | #' digest::sha1(medoids_csv) | |
146 | #' digest::sha1(medoids_bin) | |
147 | #' digest::sha1(medoids_db) | |
1c6f223e | 148 | #' } |
1c6f223e | 149 | #' @export |
d9bb53c5 | 150 | claws <- function(getSeries, K1, K2, nb_series_per_chunk, nb_items_clust1=7*K1, |
0486fbad | 151 | algoClust1=function(data,K) cluster::pam(t(data),K,diss=FALSE)$id.med, |
9f05a4a0 | 152 | algoClust2=function(dists,K) cluster::pam(dists,K,diss=TRUE)$id.med, |
a52836b2 | 153 | wav_filt="d8", contrib_type="absolute", WER="end", nvoice=4, random=TRUE, |
d9bb53c5 BA |
154 | ntasks=1, ncores_tasks=1, ncores_clust=4, sep=",", nbytes=4, |
155 | endian=.Platform$endian, verbose=FALSE, parll=TRUE) | |
ac1d4231 | 156 | { |
0e2dce80 | 157 | # Check/transform arguments |
492cd9e7 BA |
158 | if (!is.matrix(getSeries) && !bigmemory::is.big.matrix(getSeries) |
159 | && !is.function(getSeries) | |
160 | && !methods::is(getSeries,"connection") && !is.character(getSeries)) | |
0e2dce80 | 161 | { |
492cd9e7 | 162 | stop("'getSeries': [big]matrix, function, file or valid connection (no NA)") |
5c652979 | 163 | } |
eef6f6c9 BA |
164 | K1 <- .toInteger(K1, function(x) x>=2) |
165 | K2 <- .toInteger(K2, function(x) x>=2) | |
37c82bba BA |
166 | nb_series_per_chunk <- .toInteger(nb_series_per_chunk, function(x) x>=1) |
167 | # K1 (number of clusters at step 1) cannot exceed nb_series_per_chunk, because we will need | |
168 | # to load K1 series in memory for clustering stage 2. | |
169 | if (K1 > nb_series_per_chunk) | |
170 | stop("'K1' cannot exceed 'nb_series_per_chunk'") | |
eef6f6c9 BA |
171 | nb_items_clust1 <- .toInteger(nb_items_clust1, function(x) x>K1) |
172 | random <- .toLogical(random) | |
0486fbad BA |
173 | tryCatch( {ignored <- wavelets::wt.filter(wav_filt)}, |
174 | error = function(e) stop("Invalid wavelet filter; see ?wavelets::wt.filter") ) | |
eef6f6c9 BA |
175 | ctypes = c("relative","absolute","logit") |
176 | contrib_type = ctypes[ pmatch(contrib_type,ctypes) ] | |
177 | if (is.na(contrib_type)) | |
178 | stop("'contrib_type' in {'relative','absolute','logit'}") | |
7f0781b7 | 179 | if (WER!="end" && WER!="mix") |
eef6f6c9 BA |
180 | stop("'WER': in {'end','mix'}") |
181 | random <- .toLogical(random) | |
182 | ntasks <- .toInteger(ntasks, function(x) x>=1) | |
183 | ncores_tasks <- .toInteger(ncores_tasks, function(x) x>=1) | |
184 | ncores_clust <- .toInteger(ncores_clust, function(x) x>=1) | |
56857861 BA |
185 | if (!is.character(sep)) |
186 | stop("'sep': character") | |
eef6f6c9 BA |
187 | nbytes <- .toInteger(nbytes, function(x) x==4 || x==8) |
188 | verbose <- .toLogical(verbose) | |
189 | parll <- .toLogical(parll) | |
56857861 | 190 | |
2b9f5356 BA |
191 | # Binarize series if getSeries is not a function; the aim is to always use a function, |
192 | # to uniformize treatments. An equally good alternative would be to use a file-backed | |
d9bb53c5 | 193 | # bigmemory::big.matrix, but it would break the "all-is-function" pattern. |
56857861 BA |
194 | if (!is.function(getSeries)) |
195 | { | |
4bcfdbee | 196 | if (verbose) |
a52836b2 BA |
197 | cat("...Serialize time-series (or retrieve past binary file)\n") |
198 | series_file = ".series.bin" | |
199 | if (!file.exists(series_file)) | |
200 | binarize(getSeries, series_file, nb_series_per_chunk, sep, nbytes, endian) | |
4bcfdbee | 201 | getSeries = function(inds) getDataInFile(inds, series_file, nbytes, endian) |
56857861 | 202 | } |
ac1d4231 | 203 | |
95b5c2e6 | 204 | # Serialize all computed wavelets contributions into a file |
a52836b2 | 205 | contribs_file = ".contribs.bin" |
7f0781b7 | 206 | index = 1 |
cea14f3a | 207 | nb_curves = 0 |
4bcfdbee | 208 | if (verbose) |
a52836b2 BA |
209 | cat("...Compute contributions and serialize them (or retrieve past binary file)\n") |
210 | if (!file.exists(contribs_file)) | |
211 | { | |
212 | nb_curves = binarizeTransform(getSeries, | |
213 | function(series) curvesToContribs(series, wav_filt, contrib_type), | |
214 | contribs_file, nb_series_per_chunk, nbytes, endian) | |
215 | } | |
216 | else | |
217 | { | |
218 | # TODO: duplicate from getDataInFile() in de_serialize.R | |
219 | contribs_size = file.info(contribs_file)$size #number of bytes in the file | |
220 | contrib_length = readBin(contribs_file, "integer", n=1, size=8, endian=endian) | |
221 | nb_curves = (contribs_size-8) / (nbytes*contrib_length) | |
222 | } | |
4bcfdbee | 223 | getContribs = function(indices) getDataInFile(indices, contribs_file, nbytes, endian) |
8e6accca | 224 | |
2b9f5356 | 225 | # A few sanity checks: do not continue if too few data available. |
eef6f6c9 BA |
226 | if (nb_curves < K2) |
227 | stop("Not enough data: less series than final number of clusters") | |
5c652979 | 228 | nb_series_per_task = round(nb_curves / ntasks) |
eef6f6c9 BA |
229 | if (nb_series_per_task < K2) |
230 | stop("Too many tasks: less series in one task than final number of clusters") | |
ac1d4231 | 231 | |
d9bb53c5 BA |
232 | # Generate a random permutation of 1:N (if random==TRUE); |
233 | # otherwise just use arrival (storage) order. | |
2b9f5356 BA |
234 | indices_all = if (random) sample(nb_curves) else seq_len(nb_curves) |
235 | # Split (all) indices into ntasks groups of ~same size | |
236 | indices_tasks = lapply(seq_len(ntasks), function(i) { | |
237 | upper_bound = ifelse( i<ntasks, min(nb_series_per_task*i,nb_curves), nb_curves ) | |
238 | indices_all[((i-1)*nb_series_per_task+1):upper_bound] | |
239 | }) | |
240 | ||
241 | if (parll && ntasks>1) | |
242 | { | |
243 | # Initialize parallel runs: outfile="" allow to output verbose traces in the console | |
244 | # under Linux. All necessary variables are passed to the workers. | |
245 | cl = parallel::makeCluster(ncores_tasks, outfile="") | |
0486fbad | 246 | varlist = c("getSeries","getContribs","K1","K2","algoClust1","algoClust2", |
d9bb53c5 | 247 | "nb_series_per_chunk","nb_items_clust1","ncores_clust", |
a52836b2 | 248 | "nvoice","sep","nbytes","endian","verbose","parll") |
d9bb53c5 | 249 | if (WER=="mix" && ntasks>1) |
2b9f5356 BA |
250 | varlist = c(varlist, "medoids_file") |
251 | parallel::clusterExport(cl, varlist, envir = environment()) | |
252 | } | |
253 | ||
254 | # This function achieves one complete clustering task, divided in stage 1 + stage 2. | |
255 | # stage 1: n indices --> clusteringTask1(...) --> K1 medoids | |
256 | # stage 2: K1 medoids --> clusteringTask2(...) --> K2 medoids, | |
257 | # where n = N / ntasks, N being the total number of curves. | |
492cd9e7 BA |
258 | runTwoStepClustering = function(inds) |
259 | { | |
d9bb53c5 | 260 | # When running in parallel, the environment is blank: we need to load the required |
2b9f5356 | 261 | # packages, and pass useful variables. |
bf5c0844 | 262 | if (parll && ntasks>1) |
492cd9e7 BA |
263 | require("epclust", quietly=TRUE) |
264 | indices_medoids = clusteringTask1( | |
0486fbad | 265 | inds, getContribs, K1, algoClust1, nb_series_per_chunk, ncores_clust, verbose, parll) |
d9bb53c5 | 266 | if (WER=="mix" && ntasks>1) |
56857861 | 267 | { |
d9bb53c5 | 268 | if (parll) |
eef6f6c9 | 269 | require("bigmemory", quietly=TRUE) |
bf5c0844 | 270 | medoids1 = bigmemory::as.big.matrix( getSeries(indices_medoids) ) |
0486fbad | 271 | medoids2 = clusteringTask2(medoids1, K2, algoClust2, getSeries, nb_curves, |
a52836b2 | 272 | nb_series_per_chunk, nvoice, nbytes, endian, ncores_clust, verbose, parll) |
2b9f5356 | 273 | binarize(medoids2, medoids_file, nb_series_per_chunk, sep, nbytes, endian) |
56857861 BA |
274 | return (vector("integer",0)) |
275 | } | |
276 | indices_medoids | |
492cd9e7 BA |
277 | } |
278 | ||
2b9f5356 BA |
279 | # Synchrones (medoids) need to be stored only if WER=="mix"; indeed in this case, every |
280 | # task output is a set of new (medoids) curves. If WER=="end" however, output is just a | |
281 | # set of indices, representing some initial series. | |
d9bb53c5 BA |
282 | if (WER=="mix" && ntasks>1) |
283 | {medoids_file = ".medoids.bin" ; unlink(medoids_file)} | |
2b9f5356 | 284 | |
c45fd663 | 285 | if (verbose) |
e161499b BA |
286 | { |
287 | message = paste("...Run ",ntasks," x stage 1", sep="") | |
288 | if (WER=="mix") | |
289 | message = paste(message," + stage 2", sep="") | |
290 | cat(paste(message,"\n", sep="")) | |
291 | } | |
c45fd663 | 292 | |
2b9f5356 | 293 | # As explained above, indices will be assigned to ntasks*K1 medoids indices [if WER=="end"], |
d9bb53c5 | 294 | # or nothing (empty vector) if WER=="mix"; in this case, synchrones are stored in a file. |
eef6f6c9 BA |
295 | indices <- |
296 | if (parll && ntasks>1) | |
297 | unlist( parallel::parLapply(cl, indices_tasks, runTwoStepClustering) ) | |
298 | else | |
299 | unlist( lapply(indices_tasks, runTwoStepClustering) ) | |
bf5c0844 | 300 | if (parll && ntasks>1) |
492cd9e7 | 301 | parallel::stopCluster(cl) |
3465b246 | 302 | |
2b9f5356 | 303 | # Right before the final stage, two situations are possible: |
d9bb53c5 | 304 | # a. data to be processed now sit in a binary format in medoids_file (if WER=="mix") |
2b9f5356 BA |
305 | # b. data still is the initial set of curves, referenced by the ntasks*K1 indices |
306 | # So, the function getSeries() will potentially change. However, computeSynchrones() | |
307 | # requires a function retrieving the initial series. Thus, the next line saves future | |
308 | # conditional instructions. | |
8702eb86 | 309 | getRefSeries = getSeries |
2b9f5356 | 310 | |
d9bb53c5 | 311 | if (WER=="mix" && ntasks>1) |
e205f218 BA |
312 | { |
313 | indices = seq_len(ntasks*K2) | |
bccecb19 BA |
314 | # Now series (synchrones) must be retrieved from medoids_file |
315 | getSeries = function(inds) getDataInFile(inds, medoids_file, nbytes, endian) | |
2b9f5356 | 316 | # Contributions must be re-computed |
4bcfdbee | 317 | unlink(contribs_file) |
e205f218 | 318 | index = 1 |
4bcfdbee BA |
319 | if (verbose) |
320 | cat("...Serialize contributions computed on synchrones\n") | |
492cd9e7 | 321 | ignored = binarizeTransform(getSeries, |
9f05a4a0 | 322 | function(series) curvesToContribs(series, wav_filt, contrib_type), |
492cd9e7 | 323 | contribs_file, nb_series_per_chunk, nbytes, endian) |
e205f218 | 324 | } |
0e2dce80 BA |
325 | |
326 | # Run step2 on resulting indices or series (from file) | |
4bcfdbee BA |
327 | if (verbose) |
328 | cat("...Run final // stage 1 + stage 2\n") | |
0486fbad BA |
329 | indices_medoids = clusteringTask1(indices, getContribs, K1, algoClust1, |
330 | nb_series_per_chunk, ncores_tasks*ncores_clust, verbose, parll) | |
bf5c0844 | 331 | medoids1 = bigmemory::as.big.matrix( getSeries(indices_medoids) ) |
0486fbad | 332 | medoids2 = clusteringTask2(medoids1, K2, algoClust2, getRefSeries, nb_curves, |
a52836b2 | 333 | nb_series_per_chunk, nvoice, nbytes, endian, ncores_tasks*ncores_clust, verbose, parll) |
4bcfdbee | 334 | |
d9bb53c5 BA |
335 | # Cleanup: remove temporary binary files |
336 | tryCatch( | |
337 | {unlink(series_file); unlink(contribs_file); unlink(medoids_file)}, | |
338 | error = function(e) {}) | |
4bcfdbee | 339 | |
2b9f5356 BA |
340 | # Return medoids as a standard matrix, since K2 series have to fit in RAM |
341 | # (clustering algorithm 1 takes K1 > K2 of them as input) | |
eef6f6c9 | 342 | medoids2[,] |
56857861 BA |
343 | } |
344 | ||
4bcfdbee BA |
345 | #' curvesToContribs |
346 | #' | |
347 | #' Compute the discrete wavelet coefficients for each series, and aggregate them in | |
348 | #' energy contribution across scales as described in https://arxiv.org/abs/1101.4744v2 | |
349 | #' | |
eef6f6c9 | 350 | #' @param series [big.]matrix of series (in columns), of size L x n |
4bcfdbee BA |
351 | #' @inheritParams claws |
352 | #' | |
eef6f6c9 | 353 | #' @return A [big.]matrix of size log(L) x n containing contributions in columns |
4bcfdbee BA |
354 | #' |
355 | #' @export | |
0fe757f7 | 356 | curvesToContribs = function(series, wav_filt, contrib_type, coin=FALSE) |
56857861 | 357 | { |
9f05a4a0 | 358 | series = as.matrix(series) #1D serie could occur |
eef6f6c9 | 359 | L = nrow(series) |
56857861 | 360 | D = ceiling( log2(L) ) |
d9bb53c5 | 361 | # Series are interpolated to all have length 2^D |
56857861 | 362 | nb_sample_points = 2^D |
eef6f6c9 | 363 | apply(series, 2, function(x) { |
56857861 | 364 | interpolated_curve = spline(1:L, x, n=nb_sample_points)$y |
9f05a4a0 | 365 | W = wavelets::dwt(interpolated_curve, filter=wav_filt, D)@W |
d9bb53c5 | 366 | # Compute the sum of squared discrete wavelet coefficients, for each scale |
4bcfdbee | 367 | nrj = rev( sapply( W, function(v) ( sqrt( sum(v^2) ) ) ) ) |
eef6f6c9 BA |
368 | if (contrib_type!="absolute") |
369 | nrj = nrj / sum(nrj) | |
370 | if (contrib_type=="logit") | |
371 | nrj = - log(1 - nrj) | |
372 | nrj | |
373 | }) | |
56857861 BA |
374 | } |
375 | ||
492cd9e7 | 376 | # Check integer arguments with functional conditions |
56857861 BA |
377 | .toInteger <- function(x, condition) |
378 | { | |
eef6f6c9 BA |
379 | errWarn <- function(ignored) |
380 | paste("Cannot convert argument' ",substitute(x),"' to integer", sep="") | |
56857861 | 381 | if (!is.integer(x)) |
eef6f6c9 BA |
382 | tryCatch({x = as.integer(x)[1]; if (is.na(x)) stop()}, |
383 | warning = errWarn, error = errWarn) | |
56857861 | 384 | if (!condition(x)) |
eef6f6c9 BA |
385 | { |
386 | stop(paste("Argument '",substitute(x), | |
387 | "' does not verify condition ",body(condition), sep="")) | |
388 | } | |
389 | x | |
390 | } | |
391 | ||
392 | # Check logical arguments | |
393 | .toLogical <- function(x) | |
394 | { | |
395 | errWarn <- function(ignored) | |
396 | paste("Cannot convert argument' ",substitute(x),"' to logical", sep="") | |
397 | if (!is.logical(x)) | |
398 | tryCatch({x = as.logical(x)[1]; if (is.na(x)) stop()}, | |
399 | warning = errWarn, error = errWarn) | |
56857861 | 400 | x |
cea14f3a | 401 | } |