ghost
commented
8 years ago In some samples, PNN doesn't yield any prediction, so PNN returns NA. If you check following part of the code, you can realize this:
Predict the category of a new observation
@param nn A probabilistic neural network already trained
@param X A vector describing a new observation
guess.category <- function(nn, X) { probs <- guess.probabilities.of.each.category(nn, X) if(is.na(probs[1])) return(NA)
Hello
I am using pnn to predict a True / False output, but sometimes I see that the function
guess (fit, test)
returns the value NA. For me, it does not make sense at all (even in worst case, given a random value between True / False is better than returns N/A). Why and when the guess function will return NA value?
Thank you very much,