Algorithm | Tuned hyper-parameters |
---|---|
NB | priors = [0.357, 0.643]a |
SVM | C = 230; γG = 1.3 |
DT | min_samples_split = 7; min_samples_leaf = 1; max_depth = 21 |
RF | n_estimators = 106; min_samples_split = 2; min_samples_leaf = 1; max_features = 2; max_depth = 25 |
GB | n_estimators = 167; min_samples_split = 7; min_samples_leaf = 4; max_depth = 5; learning_rate = 0.5 |