基于LDA模型的药品在线评论情感分析研究

基于LDA模型的药品在线评论情感分析研究学生姓名：弋佳悦指导教师：张振芳摘要：中国互联网络信息中心(CNNIC)第51次《中国互联网络发展状况统计报告》发布，截至2022年12月，我国网民规模达10.67亿，互联网普及率达75.6%。在互联网发展的大环境下，电商平台出现了数以万计的评论数据，对挖掘线上药品消费者的需求和提升电商药品的口碑具有重要价值。文章爬取了京东电商平台上某肠胃药品的评论数据，并基于预处理以后的文本数据进行了情感倾向性分析和LD主题模型分析。结果表明，该肠胃药品具有价格偏贵、售后服务差、物流运输差、部分包装损坏等不足之处。根据所得结论对电商药品消费者提供一些决策参考，同时为该肠胃药品线上销售的升级提供一定的参考依据。关键词：电商商品评论；情感分析：药品在线评论：LDA主题模型1Research on online sentiment analysis of drug reviews based onLDA modelAbstract:The 51st Statistical Report on Internet Development in China was released by theChina Internet Network Information Center(CNNIC),and as of December 2022,the size ofChina's Internet users reached 1.067 billion,and the Internet penetration rate reached 75.6%.Inthe context of Internet development,tens of thousands of review data have appeared one-commerce platforms,which are of great value for mining user demand and improving productreputation.This paper crawls the review data of a gastrointestinal drug on the Jingdonge-commerce platform,and analyzes the emotional tendency analysis and LDA theme modelbased on the preprocessed text data.The results show that the gastrointestinal drug has theadvantages of fast curative effect,high cost performance and comprehensive therapeutic function,but it also has shortcomings such as expensive price,poor after-sales service,poor logistics andtransportation,and damage to some packaging.The conclusion provided a certain reference forthe upgrade of the online sales of gastrointestinal drugs.Key words:Martial Arts Electives;Motivation;Factors;Countermeasures2目录基于DA模型的药品在线评论情感分析研究学生姓名：弋佳悦指导教师：张振芳1引言-1.1研究背景11.2研究意义-11.3国内外文献综述1.4研究目标与研究内容2文本挖掘与分析52.1文本挖掘52.2文本预处理62.2.1数据清洗62.2.2文本分词-62.2.2去停用词和词性标注2.2.3词频统计3倾向化分析3.1基于机器学习的方法93.1.1支持向量机模型-93.1.2情感分析结果94DA主题模型-104.1LDA模型步骤及原理104.1.1LDA模型原理及步骤104.1.2寻优图4.1.3输出主题114.1.3主题数为10时LDA可视化115结论与展望---135.1对医药电商零售商的建议135.2对医药购买者的建议135.3局限性14参考文献g