projects / epclust.git / blame
| Commit | Line | Data |
|---|---|---|
| ad642dc6 BA |
1 | ## File : 03_compute-sum-of-classes_2010.r |
| 2 | ## Description : Calculer les synchrones pour chaque groupe obtenu par le | |
| 3 | ## clustering. | |
| 4 | ||
| 5 | rm(list = ls()) | |
| 6 | ||
| 7 | setwd("~/Documents/projects/2014_EDF-Orsay-Lyon2/codes/") | |
| 8 | ||
| 9 | ## 1. Read auxiliar data files #### | |
| 10 | ||
| 11 | identifiants <- read.table("identifs.txt")[ ,1] | |
| 12 | dates0 <- read.table("datesall.txt")[, 1] | |
| 13 | dates <- dates0[grep("2009", dates0)] | |
| 14 | rm(dates0) | |
| 15 | ||
| 16 | n <- length(identifiants) | |
| 17 | p <- length(dates) | |
| 18 | ||
| 19 | blocks <- c(rep(6500, 3), 5511) | |
| 20 | ||
| 21 | # Fit of the clustering : clfit | |
| 22 | #load('~/Documents/projects/2014_EDF-Orsay-Lyon2/res/clfit500.Rdata') | |
| 23 | load('~/Documents/projects/2014_EDF-Orsay-Lyon2/res/clfit200.Rdata') | |
| 24 | #load('~/Documents/projects/2014_EDF-Orsay-Lyon2/res/clfit200RC.Rdata') | |
| 25 | # table(clfit$clustering) | |
| 26 | ||
| 27 | dfclust <- data.frame(cluster = clfit$clustering) # real clustering | |
| 28 | K <- nrow(clfit$clusinfo) | |
| 29 | ||
| 30 | #dfclust2 <- dfclust | |
| 31 | dfclust$cluster <- sample(unlist(Map(rep, 1:200, table(dfclust)))) # random clust | |
| 32 | ||
| 33 | ||
| 34 | ||
| 35 | #dfclust <- head(dfclust, 50) # just for testing purpouses | |
| 36 | synchros <- matrix(0, ncol = p, nrow = K) | |
| 37 | rm(clfit) | |
| 38 | ||
| 39 | ## 2. Process the large file #### | |
| 40 | ||
| 41 | close(con) | |
| 42 | con <- file("~/tmp/2010_full.txt") # Establish a connection to the file | |
| 43 | open(con, "r") # Open the connection | |
| 44 | ||
| 45 | for(b in seq_along(blocks)){ # Reading loop | |
| 46 | nb <- blocks[b] | |
| 47 | actual <- readLines(con = con, n = nb ) | |
| 48 | auxmat <- matrix(unlist(strsplit(actual, " ")), ncol = p + 1, byrow = TRUE) | |
| 49 | rm(actual) | |
| 50 | ||
| 51 | datamat <- t(apply(auxmat[, -1], 1, as.numeric)) | |
| 52 | rownames(datamat) <- substr(auxmat[, 1], 2, 7) | |
| 53 | rm(auxmat) | |
| 54 | ||
| 55 | # obtain for each line of datamat the cluster membership (if any) | |
| 56 | clustfactor <- dfclust$cluster[match(rownames(datamat), rownames(dfclust))] | |
| 57 | ||
| 58 | # if(length(which(rownames(datamat) == "218095"))>0) print(b); | |
| 59 | ||
| 60 | for(k in 1:K){ | |
| 61 | clustk <- which(clustfactor == k) | |
| 62 | if(length(clustk) > 0) { | |
| 63 | ||
| 64 | if(length(clustk) > 1) { | |
| 65 | synchrosk <- colSums(datamat[which(clustfactor == k), ]) | |
| 66 | } else { | |
| 67 | synchrosk <- datamat[which(clustfactor == k), ] | |
| 68 | } | |
| 69 | synchros[k, ] <- synchros[k, ] + synchrosk | |
| 70 | rm(synchrosk) | |
| 71 | } | |
| 72 | } | |
| 73 | } | |
| 74 | ||
| 75 | close(con) # close connection to the file | |
| 76 | ||
| 77 | synchros <- data.frame(t(synchros), total = colSums(synchros)) | |
| 78 | ||
| 79 | write.table(synchros, file = "~/tmp/2009_synchros200RANDOM.txt") | |
| 80 | write.table(dfclust, file = "../res/dfclust200RANDOM.txt") | |
| 81 | #write.table(synchros, file = "~/tmp/2009_synchros200RC.txt") | |
| 82 |