result of using entropy
as criterion:
result of using gini
as criterion:
as shown above, using entropy will generate a over-fitting tree.
accuracy: 0.7959183673469388
classfication report:
precision recall f1-score support
0 0.88 0.88 0.88 376
1 0.31 0.32 0.32 65
avg / total 0.80 0.80 0.80 441
confusion matrix: [[330 46]
[ 44 21]]