sklearn的Adaboost Forecast_proba如何在内部运作? [英] how does sklearn's Adaboost predict_proba works internally?

问题描述

我正在使用sklearn的'predict_proba()'来预测Adaboost分类器中每个估计量属于某个类别的样本的概率.

I'm using sklearn's 'predict_proba()' to predict the probability of a sample belonging to a category for each estimators in Adaboost classifier.

from sklearn.ensemble import AdaBoostClassifier
clf = AdaBoostClassifier(n_estimators=50)
for estimator in clf.estimators_:
    print estimator.predict_proba(X_test)

Adaboost像这样实现自己的Forecast_proba():

Adaboost implements its predict_proba() like this:

https://github.com com/scikit-learn/scikit-learn/blob/bb39b49/sklearn/ensemble/weight_boosting.py#L733

DecisionTreeClassifier是sklearn用于Adaboost分类器的基本估计器. DecisionTreeClassifier如下实现其predict_proba():

DecisionTreeClassifier is sklearn's base estimator for Adaboost classifier. DecisionTreeClassifier implements its predict_proba() like this:

https://github.com. com/scikit-learn/scikit-learn/blob/bb39b49/sklearn/tree/tree.py#L549

有人告诉我Adaboost的predict_proba()是如何内部计算该概率的吗?是否有相同主题的参考文献可以帮助我?请通知我.预先感谢.

Anyone kindly tell me how predict_proba() of Adaboost internally calculates the probability? Is there any references with the same topic which can help me ? Please inform me. Thanks in advance.

推荐答案

也许是 查看全文