Package caret issue: cannot allocate vector of size

Question

I am experiencing an issue with a model I am trying to run using Caret package. It is working absolutely fine using naive Bayes function directly:

category_model <- naiveBayes(Y ~ X1 + X2 + X3, data=train_category)
predict <- predict(category_model,test_category, type = 'class')

and then using the confusion matrix to look the accuracy result.

But when I try to do the same with caret applying cross-validation:

cv_3_folds <- createMultiFolds(train_label, k = 3, times = 3)

ctrl_3 <- trainControl(method = "repeatedcv", number = 3, repeats = 3, index = cv_3_folds)

naivebayes_cv3 <- train(Y~ X1 + X2,
                        data=train_category, 
                        method = "nb", 
                        trControl = ctrl_3)

It gives me this issue: Cannot allocate vector of size 21 Go. This does not make sense this my dataset is only 500 Mo.

Its working fine with the first function so I am guessing it comes from the caret package but I don't know exactly where.

Anyone can help me on this? Thanks. M.

Package caret issue: cannot allocate vector of size

0 Answers