projects / epclust.git / blame
| Commit | Line | Data |
|---|---|---|
| ad642dc6 BA |
1 | ## File: extract-features.r |
| 2 | ||
| 3 | rm(list = ls()) | |
| 4 | ||
| 5 | ## a. Load data & libraries #### | |
| 6 | ||
| 7 | library(cluster) | |
| 8 | MOJARRITA <- Sys.info()[4] == "mojarrita" | |
| 9 | ||
| 10 | #source('http://eric.univ-lyon2.fr/~jcugliari/codes/functional-clustering.r') | |
| 11 | source('01_StBr.r') | |
| 12 | ||
| 13 | if(MOJARRITA){ | |
| 14 | setwd("~/Documents/projects/2014_EDF-Orsay-Lyon2/codes/") | |
| 15 | } else { | |
| 16 | setwd("~/ownCloud/projects/2014_EDF-Orsay-Lyon2/codes/") | |
| 17 | } | |
| 18 | ||
| 19 | ||
| 20 | matcontrib0 <- read.table(file = "~/tmp/2009_contrib.txt") | |
| 21 | n <- nrow(matcontrib0) | |
| 22 | ||
| 23 | sdcontrib <- apply(matcontrib0, 1, sd) | |
| 24 | lims <- quantile(sdcontrib, probs = c(.005, .995)) # obtain 1%-extreme data | |
| 25 | is_normal <- which((sdcontrib > lims[1]) & (sdcontrib < lims[2])) | |
| 26 | ||
| 572d139a | 27 | matcontri_ext <- matcontrib0[-is_normal, ]#"" |
| ad642dc6 BA |
28 | matcontrib <- matcontrib0[is_normal, ] # wipe out aberrant data |
| 29 | ||
| 30 | matcontrib <- t(apply(matcontrib, 1, function(x) x / sum(x))) | |
| 31 | matcontrib <- t(apply(matcontrib, 1, function(p) log(p / (1 - p)) )) | |
| 32 | ||
| 33 | ||
| 34 | ## b. Transform data & compute CI #### | |
| 35 | ci <- CI(matcontrib) | |
| 36 | tdata <- ci$tdata; rownames(tdata) <- rownames(matcontrib) | |
| 37 | selvar <- ci$selectv | |
| 38 | ||
| 39 | ## c. Clustering ########## | |
| 40 | K <- 200 | |
| 41 | system.time( | |
| 572d139a BA |
42 | |
| 43 | #TODO: cette partie en C | |
| 44 | ||
| ad642dc6 BA |
45 | clfit <- clara(x = tdata[, selvar], |
| 46 | k = K, | |
| 47 | sampsize = 4000, | |
| 48 | samples = 4, | |
| 49 | rngR = TRUE) | |
| 50 | ) | |
| 51 | ||
| 52 | #save(clfit, file = 'clfit500.Rdata') | |
| 53 | # save(clfit, file = 'clfit200RC.Rdata') | |
| 54 | save(clfit, file = 'clfit200.Rdata') | |
| 55 | ||
| 56 | rm(ci, matcontrib0, is_normal, lims, selvar) | |
| 57 | gc() | |
| 58 | ||
| 59 | ## d. Analyze results ########## | |
| 60 | ||
| 61 | #1. Répartition du nombre d'observation par cluster | |
| 62 | #plot(sort(table(clfit$clustering), decreasing = TRUE), | |
| 63 | # type = 'l', ylab = 'Fréquence', xlab = 'Classe') | |
| 64 | ||
| 65 | ||
| 66 | #clust <- res$clustering | |
| 67 | # centres <- aggregate(conso, clust) | |
| 68 | # table(clust) | |
| 69 | ||
| 70 | #sel_veille <- as.Date(rownames(conso)[sel - 1]) | |
| 71 | #sel_lendemain <- as.Date(rownames(conso)[sel + 1]) | |
| 72 | ||
| 73 | #res_clust <- data.frame(date = rownames(conso), | |
| 74 | #veille = weekdays(sel_veille), | |
| 75 | #lendemain = weekdays(sel_lendemain), | |
| 76 | # clust = clust) | |
| 77 | ||
| 78 | #for(k in 1:K) { | |
| 79 | # assign(paste0("dates_clust", K), | |
| 80 | # substr(subset(res_clust, clust == k)$date, 1, 7) ) | |
| 81 | #} | |
| 82 | ||
| 83 | #dev.off() | |
| 84 | ||
| 85 | #save(file = paste0(dtitle, "_clust.Rdata"), | |
| 86 | #res_clust, selvar, K, gap) | |
| 87 | #} | |
| 88 | ||
| 89 | #dates_clust1 <- substr(subset(dates, clust == 1)$date, 1, 7) |