基于多网站的商品评论倾向性研究:以手机为例  

Sentiment Orientation Analysis of Product Reviews among Multiple Websites : Taking the Mobile Phone as an Example

作　　者：宋恩梅[1] 何帆[1] 

机构地区：[1]武汉大学信息管理学院

出　　处：《图书馆学研究》2016年第2期85-92,共8页Research on Library Science

基　　金：国家社会科学基金项目"参与协作式情报空间的构建与应用研究"(10CTQ016)的研究成果之一

摘　　要：网上购物已成为社会生活中的一部分,商品购买评论为消费者和厂商提供了重要的产品信息反馈。通过构建领域本体、改进情感词典并构造商品特征匹配组的方法,比较京东网、亚马逊商城、天猫商城三个购物网站的商品评论情感倾向。结果发现不同网站关于手机各部件特征的评论在数量上和情感倾向上差异明显,一定程度上反映出网站用户群体及商品版本之间的差异。通过比较多网站的商品评论,帮助消费者获得较单一网站更为可信的商品评论,并为厂商改进产品提供参考。Since online shopping has been an important part of our social life, the product reviews provide vital information about the commodity for consumers and manufactures. Using the methods of building the product domain ontology,improving the sentiment lexicon and constructing the product feature-sentiment match,this paper analyzes the sentiment orientation of the reviews on JD, Amazon and Tmall. The result turns out that the reviews about product component features on the three websites have significant difference in quantity and sentiment orientation,and it shows the differences in website user groups and in versions of the product to a certain extent. The comparative analysis among multiple websites can help the consumers get more authentic product reviews than getting them from a single website,and the manufactures can get the reference on improving the product.

关 键 词：商品评论 情感倾向 多网站 本体构建 情感词典 

分 类 号：F724.6[经济管理—产业经济] G358[文化科学—情报学]