I have used Random Forest from scikit-learn for a classification task:
rf = ensemble.RandomForestClassifier(n_estimators=1000,
max_depth = 10,
max_features = 2,
max_samples = 0.5)
rf.fit(xtrain, ytrain)
train_output_rf = rf.predict(xtrain)
test_output_rf = rf.predict(xtest)
In practice, the trained model labels the test sample with a single output (class). Is there any way or alternative algorithm to achieve a list of probable classes for the test sample instead of single output (class)? For instance:
1. A (80%)
2. B (15%)
3. C (4%)
4. D (1%)
Aucun commentaire:
Enregistrer un commentaire