scikit-learn 分类模型各种评估指标使用

import pandas as pd
from sklearn import metrics


# ModelEvolution
class ModelEvolution(object):
    def __init__(self):
        pass

    @staticmethod
    def report(test_y, score, pred_y):  # test_y 原始标签, score 模型输出概率值, pred_y 预测值
        if len(test_y) != len(score):
            print("The Lengths of test set and predicted set are not match!")
            exit(-1)

        measureNames, measureValues = [], []  # measureNames 评价指标的名称, measureValues 对应的值
        indicators = pd.DataFrame()

        accu = ModelEvolution.accuracy(test_y, score)
        measureNames.append("Accuracy:")
        measureValues.append(str(accu))

        prec = ModelEvolution.precision(test_y, score)
        measureNames.append('Precision:')
        measureValues.append(str(prec))

        reca = ModelEvolution.recall(test_y, score)
        measureNames.append('Recall:')
        measureValues.append(str(reca))

        f1 = ModelEvolution.F1(test_y, score)
        measureNames.append('F1:')
        measureValues.append(str(f1))

        ap = ModelEvolution.AP(test_y, pred_y)
        measureNames.append('AP:')
        measureValues.append(str(ap))

        auc = ModelEvolution.AUC(test_y, pred_y)
        measureNames.append('AUC:')
        measureValues.append(str(auc))

        cmat = ModelEvolution.CMatrix(test_y,score)
        measureNames.append('ConfusionMatrix:')
        measureValues.append(str(cmat))

        indicators['measureNames'] = measureNames
        indicators['measureValues'] = measureValues
        return indicators  # 返回评估值

    @staticmethod
    def accuracy(test_y, score):  # 准确率
        return metrics.accuracy_score(test_y, score)

    @staticmethod
    def precision(test_y, score):  # 精确率
        return metrics.precision_score(test_y, score, average=None)

    @staticmethod
    def recall(test_y, score):  # 召回率
        return metrics.recall_score(test_y, score, average=None)

    @staticmethod
    def F1(test_y, score):  # F1
        return metrics.f1_score(test_y, score, average=None)

    @staticmethod
    def AP(test_y, pred_y):  # 平均精确率
        return metrics.average_precision_score(test_y, pred_y)

    @staticmethod
    def AUC(test_y, pred_y):  # AUC
        return metrics.roc_auc_score(test_y, pred_y)

    @staticmethod
    def CMatrix(test_y, score):  # 混淆矩阵
        return metrics.confusion_matrix(test_y, score)


if __name__ == '__main__':
    pass
    # ModelEvolution.report()

未完成，稍后继续
声明： 总结学习，有问题或不当之处，可以批评指正哦，谢谢。