the training data compared to the single decision tree model we saw earlier