机器学习-分类问题

Question

在预测模型中的任何值时，是否有任何命令调用错误分类的值，简而言之就是那些分类不正确的值。

Answer 1

scikit-learn API中没有此类方法。不过，您可以自己编写几行：

import numpy as np

X_train, X_test, y_train, y_test = ...  # some dataset
model = ...  # some scikit learn model
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

# command to get misclassified examples assuming you have numpy arrays
X_test[np.argwhere(y_pred != y_test)]