转自：https://blog.csdn.net/u011630575/article/details/78594791

1.例子

from sklearn.metrics import cohen_kappa_score
y_true = [1,1,1]
y_pred = [2,2,2]
print(cohen_kappa_score(y_true, y_pred))

0.0

说明kappa系数的计算与label是有关系的，不像ARI和NMI计算与label无关：

NMI：

from sklearn.metrics.cluster import normalized_mutual_info_score
y_true = [1,1,1]
y_pred = [2,2,2]
print(normalized_mutual_info_score(y_true, y_pred))

1.0

ARI：

from sklearn.metrics.cluster import adjusted_rand_score
y_true = [1,1,1]
y_pred = [2,2,2]
print(adjusted_rand_score(y_true, y_pred))

1.0