信息熵计算（自己编写的python代码，垃圾，高手绕道）

最新推荐文章于 2019-04-18 11:36:03 发布

奋斗的小乌龟

最新推荐文章于 2019-04-18 11:36:03 发布

阅读量5.9k

点赞数

分类专栏：人工智能文章标签： python math eclipse

本文链接：https://blog.csdn.net/guo20082200/article/details/78001762

版权

人工智能专栏收录该内容

3 篇文章 0 订阅

订阅专栏

# -*- coding:utf-8 -*-
'''
Created on 2017年9月15日


@author: snow
'''
import csv;
import math;
fileName = "AllElectronics.csv";
def allData():
    csv_reader = csv.reader(open(fileName, encoding='UTF-8'));
    fileContent = [];
    for row in csv_reader:
        fileContent.append(row);
    headers = fileContent[0];
    dataContent = [];
    labels=[];
    for i in range(1,len(fileContent)):
        dataContent.append(fileContent[i][-1]);
        labels.append(fileContent[i][-1]);
        
    dataSet = [];
    for row in (dataContent):
        rowData=row[1:len(row)-1];
        dataSet.append(rowData);
    return headers,dataContent,labels,dataSet;


headers,dataContent,labels,dataSet = allData();


numEntries = len(labels);
def calEnt(labels):
    labelCounts={};
    for lable in labels:
        if lable not in labelCounts.keys():
            labelCounts[lable] = 0;
        labelCounts[lable]+=1;
    shannonEnt=0.0;
    for key in labelCounts.keys():
        print(labelCounts[key]);
        prob = float(labelCounts[key])/numEntries;
        shannonEnt -= prob * math.log(prob,2) # 以2为底的对数
    return shannonEnt


res = calEnt(labels);
print(res);