聚类评估指标系列(一)：标准化互信息NMI计算步骤及其Python实现

本文转自：标准化互信息NMI计算步骤及其Python实现

标准化互信息NMI (Normalized Mutual Information)常用在聚类评估中。

标准化互信息NMI计算步骤

Python 实现

代码：

?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

'''  利用Python实现NMI计算''' import math import numpy as np from sklearn import metrics def NMI(A,B):     # 样本点数     total = len(A)     A_ids = set(A)     B_ids = set(B)     # 互信息计算     MI = 0     eps = 1.4e-45     for idA in A_ids:         for idB in B_ids:             idAOccur = np.where(A==idA)    # 输出满足条件的元素的下标             idBOccur = np.where(B==idB)             idABOccur = np.intersect1d(idAOccur,idBOccur)   # Find the intersection of two arrays.             px = 1.0*len(idAOccur[0])/total             py = 1.0*len(idBOccur[0])/total             pxy = 1.0*len(idABOccur)/total             MI = MI + pxy*math.log(pxy/(px*py)+eps,2)     # 标准化互信息     Hx = 0     for idA in A_ids:         idAOccurCount = 1.0*len(np.where(A==idA)[0])         Hx = Hx - (idAOccurCount/total)*math.log(idAOccurCount/total+eps,2)         Hy = 0     for idB in B_ids:         idBOccurCount = 1.0*len(np.where(B==idB)[0])         Hy = Hy - (idBOccurCount/total)*math.log(idBOccurCount/total+eps,2)     MIhat = 2.0*MI/(Hx+Hy)     return MIhat     if __name__ == '__main__':     A = np.array([1,1,1,1,1,1,2,2,2,2,2,2,3,3,3,3,3])     B = np.array([1,2,1,1,1,1,1,2,2,2,2,3,1,1,3,3,3])     print(NMI(A,B))     print(metrics.normalized_mutual_info_score(A,B))   # 直接调用sklearn中的函数

运行结果：

0.3645617718571898
0.3646247961942429