I'm using the CARET package in R for multiclass classification. I have 3 classes and I use the method train for training purposes. Here is the code for it:
trained.model.rf <- train(x = dataset.train[,-c(ncol(dataset.train))], y = dataset.train[,ncol(dataset.train)],method='rf',
trControl=trainControl(method="cv",number=10), allowParallel=TRUE, tuneLength = 6)
model.rf <- trained.model.rf$finalModel
result.rf <- predict(model.rf, dataset.test, type="response")
For the dataset.train I have the three classes together.
How can I tell if this is an one-versus-one or one-versus-all aproach?