R/cv10lda2.R

In dprep: Data Pre-Processing and Visualization Functions for Classification

cv10lda2 <-
function (data) 
{
    n <- dim(data)[1]
    p <- dim(data)[2]
    salida <- matrix(0, 1, 10)
    azar <- data[rank(runif(n)), ]
    parti <- floor(n/10)
    for (j in 1:10) {
        cc <- ((j - 1) * parti + 1):(j * parti)
        if (j == 10) {
            cc <- ((j - 1) * parti + 1):n
        }
        datap <- azar[cc, ]
        datat <- azar[-cc, ]
        tempo <- MASS::lda(as.matrix(datat[, 1:p - 1]), datat[, p])
        tempo1 <- predict(tempo, as.matrix(datap[, 1:p - 1]))$class
        salida[j] <- sum(tempo1 != as.numeric(datap[, p]))
    }
    gooderr <- n - sum(salida)
    return(gooderr)
}