基于主题模型和情感分析的垃圾评论识别方法研究

发布时间：2018-04-16 18:13

本文选题：产品评论 + 垃圾评论　；参考：《计算机科学》2017年10期

【摘要】：随着电子商务的飞速发展,网络购物越来越被消费者认同,而随之产生的产品评论给消费者的购买决策带来了影响。产品评论是指用户在购物站点上对商品的评价信息,而经过分析和研究发现这些评论中充斥着大量的垃圾评论,因此垃圾评论的识别成了电子商务在提高服务质量的过程中需解决的重要问题之一。根据垃圾评论的主要特点提出LDA-SP(LDA-Sentiment Polarity)垃圾评论识别方法。首先利用LDA主题模型过滤出内容型垃圾评论,然后结合情感分析识别出欺骗型垃圾评论。对网络商城的大量评论数据进行准确度分析实验的结果表明,LDA-SP方法的识别准确度高于传统的LDA主题模型和单一的情感极性分析方法,能够有效地检测垃圾评论,从而使产品评论信息更加客观准确,为电子商务用户提供了有效的参考信息。
[Abstract]:With the rapid development of electronic commerce, online shopping is more and more recognized by consumers.Product reviews refer to the information that users evaluate goods on shopping sites, and after analysis and research, it is found that these comments are filled with a lot of junk comments.Therefore, the identification of spam reviews has become one of the important problems to be solved in the process of improving the service quality of electronic commerce.According to the main characteristics of garbage comment, this paper presents a method of garbage comment recognition based on LDA-SP(LDA-Sentiment Policy.Firstly, the LDA theme model is used to filter out the content-type spam comments, and then the deceptive spam comments are identified by affective analysis.The experimental results of accuracy analysis on a large number of comment data in network mall show that the recognition accuracy of LDA-SP method is higher than that of traditional LDA theme model and single emotional polarity analysis method, and it can effectively detect garbage comments.Thus, the product comment information is more objective and accurate, and provides effective reference information for e-commerce users.
【作者单位】：武汉理工大学计算机科学与技术学院;
【基金】：国家社会科学基金(15BGL048) 国家863计划项目(2015AA015403)资助
【分类号】：TP391.1

【相似文献】