When there are a lot of categorical features, RF can be very slow in particular with ridgeRF. In that case, it makes sense to onehot encode the data:
E.g.:
servo_onehot_translator <- onehot::onehot(servo)
servo <- as.data.frame(predict(servo_onehot_translator, servo))
When there are a lot of categorical features, RF can be very slow in particular with ridgeRF. In that case, it makes sense to onehot encode the data:
E.g.: