基于文本情感分析的电商产品评论数据研究  

Research on E-Commerce Product Comment Data Based on Text Sentiment Analysis

作　　者：章蓬伟 贾钰峰 邵小青 拜尔娜·木沙 赵裕峰 ZHANG Pengwei;JIA Yufeng;SHAO Xiaoqing;MUSA Birna;ZHAO Yufeng(College of Information Science and Engineering,Xinjiang University of Science and Technology,Korla Xinjiang 841000,China;Department of Business Administration,Xinjiang University of Science and Technology,Korla Xinjiang 841000,China;Department of Electrical Engineering,Liaoning Ecological Engineering Vocational College,Shenyang 110870,China)

机构地区：[1]新疆科技学院信息科学与工程学院,新疆库尔勒841000 [2]新疆科技学院工商管理系,新疆库尔勒841000 [3]辽宁生态工程职业学院电气工程学院,沈阳110122

出　　处：《微处理机》2020年第6期58-62,共5页Microprocessors

基　　金：新疆巴音郭楞蒙古自治州科学技术研究(星火项目)(2019018);新疆科技学院校级科研项目(2020PT027)。

摘　　要：文本情感分析是当前数据挖掘研究的热点,具有极高的应用价值。对电商在线评论进行文本情感分析与数据挖掘具有很高的学术价值和社会经济效益。针对电商在线评论的特点,采用八爪鱼数据采集器爬取京东商城评论数据,采用Jiaba分词,进行情感标注以及关键词云生成。通过Python基于情感词典进行情感倾向性分析,对情感词、程度副词、否定词进行遍历,计算评论语句的情感值,并使用SnowNLP类库对文本情感值得分进行可视化展示,以柱状图直观展示语句情感分数分布情况,很好验证了实验结论。As a hot topic in current data mining research,text sentiment analysis has high application value.It is of great academic value and social and economic benefits to conduct text sentiment analysis and data mining on E-commerce online reviews.According to the characteristics of E-commerce online reviews,Octopus data collector is used to crawl the review data of JD.COM Mall,and Jiaba word segmentation is used to mark sentiments and generate keyword cloud.Through Python's sentiment tendency analysis based on sentiment dictionary,the sentiment words,degree adverbs and negative words are traversed,and the sentiment values of comment sentences are calculated.The SnowNLP class library is used to visually display the text sentiment scores,and the histogram is used to visually display the distribution of sentiment scores of sentences,which well verifies the experimental conclusions.

关 键 词：情感分析 情感词典 SnowNLP类库 词云 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]