Optimizing the accuracy

The accuracies obtained in the previous exercise,

acc_undersample
acc_prior
acc_loss_matrix
acc_weights

are loaded in your workspace again. Have a look at them. Which tree performs best, when looking at accuracy alone?