R help - Oct 2013 - Logistic regression over LOOCV

Hello all.

I have this code:

myLOOCV <- function(myformula, data) {
    Y <- all.vars(myformula)[1]
    Scores<- numeric(length(data[,1]))
    for (i in 1:length(data[,1])) {
        train <- data[-i,]
        test <- data[i,]
        myModel <- lrm(myformula, train)
        Scores[i] <- predict(myModel, test,type="mean")
        }
    factor(Scores,labels=levels(data[,Y]))
    }

When I call the method myLOOCV, myLOOCV(value, data), I have this error:

Error in factor(Scores, labels = levels(data[, Y])) :
  invalid 'labels'; length 4 should be 1 or 180

I don't understand it! Could anyone solve it?

Thanks in advance.

	[[alternative HTML version deleted]]