如何在 Pandas 中找到类的 Jaccard 相似度

Question

我想在我的数据集中找到每对组之间的 Jaccard 相似度。我的数据如下，第一列是我的数据，第二列是class lable：

import pandas as pd
import numpy as np
df = pd.DataFrame({'Data' : ["a1","a2","a3","a4","a5","a6","a7"], 'ClassLable' :     ["c1","c2","c2","c2","c3","c3","c1"]}); df
df2 = pd.DataFrame({'Data' : ["a1","a2","a4","a6","a7","a8","a9"], 'ClassLable' : ["c11","c21","c21","c12","c13","c13","c11"]}); df2

我想计算 df 和 df2 之间每对类标签的 Jaccard 指数。例如：

c1class = pd.DataFrame({'Data':["a1","a7"]})
c11class = pd.DataFrame({'Data':["a1","a9"]})
Jaccard = 1/3

换句话说，对于 df1 和 df2，我想在每个类标签的联合上找到相交的项目

Answer 1

您是否在寻找类似的东西：

from sklearn.metrics import jaccard_similarity_score

jaccard_similarity_score(df['Data'],df2['Data'])
Out[92]: 0.2857142857142857

jaccard_similarity_score(c1class, c11class)
Out[93]: 0.5