【Python-ML】聚类的性能评价指标

参考：http://scikit-learn.org/stable/modules/clustering.html#clustering-performance-evaluation

1、兰德指数

from sklearn import metrics
labels_true = [0, 0, 0, 1, 1, 1]
labels_pred = [0, 0, 1, 1, 2, 2]

print (metrics.adjusted_rand_score(labels_true, labels_pred))


2、互信息

from sklearn import metrics
labels_true = [0, 0, 0, 1, 1, 1]
labels_pred = [0, 0, 1, 1, 2, 2]
print (metrics.adjusted_mutual_info_score(labels_true, labels_pred) )


3、Homogeneity, completeness and V-measure

同质性homogeneity：每个群集只包含单个类的成员。

完整性completeness：给定类的所有成员都分配给同一个群集。

两者的调和平均V-measure。

from sklearn import metrics
labels_true = [0, 0, 0, 1, 1, 1]
labels_pred = [0, 0, 1, 1, 2, 2]
print (metrics.homogeneity_score(labels_true, labels_pred))
print (metrics.completeness_score(labels_true, labels_pred))
print (metrics.v_measure_score(labels_true, labels_pred))

4、Fowlkes-Mallows scores



from sklearn import metrics
labels_true = [0, 0, 0, 1, 1, 1]
labels_pred = [0, 0, 1, 1, 2, 2]
print (metrics.fowlkes_mallows_score(labels_true, labels_pred))

5、Silhouette Coefficient 轮廓系数
参考：
http://blog.csdn.net/fjssharpsword/article/details/79161570

6、Calinski-Harabaz Index

类别内部数据的协方差越小越好，类别之间的协方差越大越好，这样的Calinski-Harabasz分数会高。