AUC的含义以及Python代码手动实现

主要以二分类为例，AUC曲线是ROC曲线下方的面积。ROC曲线的横坐标是假阳率，纵坐标是真阳率。

AUC 的计算

1）将所有正样本pos1,pos2,,,posm与所有负样本neg1,neg2,negn组成m*n样本对；

2）若正样本的预测概率大于负样本的预测概率，计数count增加1；若相等，计数count增加0.5，否则增加0；

3）AUC = count/m*n;

import numpy as np
from sklearn.metrics import roc_curve
from sklearn.metrics import auc

#python 手动实现
def auc_calculate(labels,preds,n_bins=100):
    pos_len = sum(labels)
    neg_len = len(labels) - pos_len
    total_num = pos_len * neg_len
    pos_histogram = [0 for i in range(n_bins)]
    neg_histogram = [0 for i in range(n_bins)]
    bin_width = 1.0 / n_bins
    for i in range(len(labels)):
        nth_bin = int(preds[i]/bin_width)
        if labels[i]==1:
            pos_histogram[nth_bin] += 1
        else:
            neg_histogram[nth_bin] += 1
    sum_neg = 0
    count = 0
    for i in range(n_bins):
        # 对于同一个i,pos_hisg[i] 大于前面所有neg_hisg[0~i-1]的概率，所以要进行neg_hisg的累加，对于当前i,概率相等，取两者相乘再除以2
        count += (pos_histogram[i]*sum_neg + pos_histogram[i]*neg_histogram[i]*0.5)
        sum_neg += neg_histogram[i]

    return count / float(total_num)

if __name__ == '__main__':

    y = np.array([1,0,0,0,1,0,1,0,1,1,0,0,0,1,0,1,0,1])
    pred = np.array([0.9, 0.1, 0.3, 0.1,0.4,0.9,0.66,0.7,0.8, 0.1, 0.3, 0.1,0.4,0.9,0.66,0.7,0.8,0.9])
    fpr, tpr, thresholds = roc_curve(y, pred, pos_label=1)
    print("sklearn官方代码结果:",auc(fpr, tpr))
    print("自己代码结果:",auc_calculate(y,pred))

结果：

sklearn官方代码结果: 0.7250000000000001
自己代码结果: 0.725