pca.R

# principal component analysis
# removeVariables <- c("hydroMode", "hm149")
# keepVariables <- setdiff(names(train), removeVariables)
# trpc <- scale(train[, keepVariables])
# tepc <- scale(test[, keepVariables])
trpc <- scale(alltr)
tepc <- scale(allte)
pc <- princomp(trpc)
cumpc <- cumsum(pc$sd^2)/sum(pc$sd^2)
pc.tr <- as.data.frame(pc$scores[,1:14])
pc.te <- as.data.frame((tepc %*% pc$loadings))[,1:14]