1 ## File : 01_extract-features_2009.r
2 ## Description : Using the full data matrix, we extract handy features to
7 source('http://eric.univ-lyon2.fr/~jcugliari/codes/functional-clustering.r')
8 setwd("~/ownCloud/projects/2014_EDF-Orsay-Lyon2/codes/")
10 ## 1. Read auxiliar data files ####
12 identifiants <- read.table("identifs.txt")[ ,1]
13 dates0 <- read.table("datesall.txt")[, 1]
14 dates <- dates0[grep("2009", dates0)]
17 n <- length(identifiants)
20 blocks <- c(rep(6500, 3), 5511)
22 # table( substr(dates, 11, 15) ) # Sunlight time saving produces an
23 # unbalanced number of time points
24 # per time stepa across the year
27 ## 2. Process the large file ####
30 con <- file("~/tmp/2009_full.txt") # Establish a connection to the file
31 open(con, "r") # Open the connection
33 for(b in seq_along(blocks)){ # Reading loop
35 actual <- readLines(con = con, n = nb )
36 auxmat <- matrix(unlist(strsplit(actual, " ")), ncol = p + 1, byrow = TRUE)
39 datamat <- t(apply(auxmat[, -1], 1, as.numeric))
40 rownames(datamat) <- substr(auxmat[, 1], 2, 7)
43 auxDWT <- t(apply(datamat, 1, toDWT))
44 auxcontrib <- t(apply(auxDWT, 1, contrib))
48 matcontrib <- auxcontrib
50 matcontrib <- rbind(matcontrib, auxcontrib)
55 close(con) # close connection to the file
57 write.table(matcontrib, file = "~/tmp/2009_contrib.txt")