aucの計算

calc_auc.py

import numpy as np
from itertools import combinations
from sklearn.metrics import roc_auc_score

def calc_auc_slow(label, pred):
    #激遅
    n = label.shape[0]
    all_combs = np.array(list(combinations(np.arange(n), 2)))
    label_pairs = label[all_combs]
    pred_pairs = pred[all_combs]
    denom = np.bincount(label).prod()
    mole = np.logical_and(label_pairs[:, 0] != label_pairs[:, 1], pred_pairs[:, 0] == pred_pairs[:, 1]).sum() * 0.5 + \
                 np.logical_and(label_pairs[:, 0] > label_pairs[:, 1], pred_pairs[:, 0] > pred_pairs[:, 1]).sum() + \
                 np.logical_and(label_pairs[:, 0] < label_pairs[:, 1], pred_pairs[:, 0] < pred_pairs[:, 1]).sum()
    return mole/denom

if __name__ == '__main__':
    label = np.array([1, 0, 1, 0, 0, 1, 0, 0, 1])
    pred = np.array([1, 2, 3, 2, 3, 2, 2, 1, 1])
    assert(roc_auc_score(label, pred) == calc_auc_slow(label, pred))
    print(roc_auc_score(label, pred))
    print(calc_auc_slow(label, pred))