摘要
Web2.0时代,阅读在线产品评论已经成为人们购物前的一种习惯。然而,网络上的评论数量巨大且观点不一,消费者很难获取到真正对其有用的评论。本文从研究中文在线产品评论的有用性评估入手,结合中文在线评论的特点,构建了评论有用性评估特征体系。以二分类思想为中心,基于文本挖掘的基本流程,实现对中文产品评论的分类,并考察了评论内容各特征对分类效果的影响。结果表明,本文提出的评估方法能有效识别出有用评论,并且发现浅层句法特征在分类中的贡献度较高,语义特征与情感特征则会因语料类型的不同而有不同的分类贡献度。
In the web2.0 era, before going shopping, reading online product reviews has become a habit oi people. However, the number of reviews on the Internet is huge and opinions are different, consumers are difficult to gain helpful reviews. This paper studies on assessing the helpfulness of Chinese online product reviews.Combined with the characteristics of Chinese online reviews ,we construct a feature system about assessing reviews helpfulness. Taking the thought of binary classification as a center, based on the basic flow of text mining, to achieve the Chinese product reviews classification. And study on the degree of each feature of review content affecting the classification effect. The results show that the proposed evaluation method can effectively identify helpful reviews, and found shallow syntactic features has higher contribution in the classification. In addition, semantic features and sentiment features have different classification contribution on different type of corpus.
出处
《未来与发展》
2016年第3期37-43,共7页
Future and Development
关键词
在线产品评论
有用性
评估
文本挖掘
online product reviews, helpfulness, assessing, text mining
