Minepy—使用python计算最大互信息系数(MIC)

MIC 即：Maximal Information Coefficient 最大互信息系数。

https://blog.csdn.net/qtlyx/article/details/50780400

MIC可以用来衡量线性或非线性的相互关系。

 

---

算法对比

 

https://zhuanlan.zhihu.com/p/53092905

---

MIC算法可以通过python的minpy包来实现。

python 安装 minepy

安装命令：

pip install minepy

---

安装报错，提示没有C++编译器

https://www.cnblogs.com/xiaoxineryi/p/12405063.html

从Unofficial Windows Binaries for Python Extension Packages（这个网站包含了很多Python扩展包的非官方Windows二进制文件）下载对应的minepy.whl

 

安装命令

pip install whl路径

安装报错，

ERROR: minepy-1.2.4-cp38-cp38-win_amd64.whl is not a supported wheel on this platform

---

需要选择和自己Python版本对应的版本

查询python支持的whl格式

from pip._internal import pep425tags

print(pep425tags.get_supported())
#打印支持的whl版本

>>>[('cp37', 'cp37m', 'win_amd64'), ('cp37', 'none', 'win_amd64'), ('py3', 'none', 'win_amd64'), ('cp37', 'none', 'any'), ('cp3', 'none', 'any'), ('py37', 'none', 'any'), ('py3', 'none', 'any'), ('py36', 'none', 'any'), ('py35', 'none', 'any'), ('py34', 'none', 'any'), ('py33', 'none', 'any'), ('py32', 'none', 'any'), ('py31', 'none', 'any'), ('py30', 'none', 'any')]

下载适合的whl版本，安装成功

---

使用minpy

https://s0minepy0readthedocs0io.icopy.site/en/1.2.2/python.html

import numpy as np

import pandas as pd

from minepy import MINE



inpath = r'D:\Python\input_data\abc.xlsx'

outpath = r'D:\Python\output_data\abc-MIC系数.xlsx'


data_df = pd.read_excel(inpath)


depend_names = ['变量A',

                '变量B',

                '变量C']


output_mic = pd.DataFrame(columns=['变量1','变量2','MIC系数'])


for var1 in depend_names:

    for var2 in depend_names:

        x = data_df[var1]

        y = data_df[var2]

        mine = MINE(alpha = 0.6, c = 15)

        mine.compute_score(x, y)

        print(mine.mic())
   

        out_dict = {'变量1': var1,

                '变量2': var2,

                'MIC系数': mine.mic(),

                }

        output_mic = output_mic.append(out_dict, ignore_index=True)

output_mic.to_excel(outpath)