How to use k-fold cross validation in naive bayes classifier?

Question

I'm trying to classify text using naive bayes classifier, and also want to use k-fold cross validation to validate the result of classification. But I'm still confused how to use the k-fold cross validation. As i know that k-fold divide data to k subsets, then one of the k subsets is used as the test set and the other k-1 subsets are put together to form a training set. And i think as training set the data must have label to be trained. So to use k-fold cross validation the required data is the labeled data?, is it right?, and how about non labeled data?.

Typically, for any supervised learning, the data needs to be labeled. And then again, for evaluation, the data needs to be labeled. — Chthonic Project
so actually the k-fold cross validation used in naive bayes for training, and not for testing? — Muhammad Haryadi Futra Iskanda
Well you will train the Naive Bayes with with k-1 subsets. When the model is created you will evaluate the model with the remaining subset. The model will predict a class and you can compare this predicted result with the corret result. — user

mohammad karim hardani asl mohammad karim hardani asl · Accepted Answer · 2015-12-26T15:42:34

for non labeled data you must use clustering methods, for nb maybe this code would help you:

[testF, trainF] = kfolds(Features,k);
[testL, trainL] = kfolds(Label,k);
c = size(Features);
for i=1:k
    LabelTrain = trainL{i};
    LabelTest = testL{i};
    FeaturesTrain = trainF{i};
    FeaturesTest = testF{i};
    nb = NaiveBayes.fit(FeaturesTrain,LabelTrain);
    Class = predict(nb,FeaturesTest);
    predict_Class(i)=sum(Class==LabelTest);
end
predict_all = sum(predict_Class)/c(1);

kfolds function would separate your data to k folds.

cheers

How to use k-fold cross validation in naive bayes classifier?

1 Answers