摘要： 客户评论在人们的日常生活中越来越重要,人们希望从客户评论中获取商品的用户体验信息。客户评论数量的急剧增长使得用户快速、精准地获取有用的信息变得较为困难。为此,提出一种能够自动提取用户体验信息的方法。该方法通过语义片段过滤评论中的冗余信息,提取产品特征词及特征描述词,将其结合组成用户体验信息,自动获取信息能够迅速、准确地从客户评论中提取信息。实验结果证明了该方法的有效性,并且能够保证较高的准确率与查全率。

关键词: 客户评论, 特征挖掘, 情感分析, 语义片段提取, 用户体验, 语义相似度

Abstract: Customer reviews are playing an increasingly important role in people’s daily lives,from which people want to obtain some information about user experience.However,with the continuous development of the Internet,it is pretty difficult for users to get the useful information in a rapid and accurate way.The common practice is to collect experience information manually or half-manually,and calculate the frequency of tem.This paper presents an automatic method to extract information about the user experience from customer reviews,it extracts product features and feature description through semantic segment filtering redundant information,and consists of user experience information,it implements information extraction rapidly and precisely.Abundant experiments show that this method is available and can guarantee very high precision and recall ratio.

Key words: customer reviews, feature mining, emotion analysis, semantic segment extraction, user experience, semantic similarity

中图分类号: 

• TP391