论文部分内容阅读
网络用户评论信息量庞大,对用户评论内容的语义挖掘具有重要的意义。现有的研究主要是从评论中产品属性的挖掘以及用户情感的分析,缺乏语义层面的分析。本文提出一种分析方法,实现用户评论的语义挖掘。首先将网络用户评论抽象为“实体-属性-值”的形式,并对评论信息进行断句、分词,形成具有语义特征的评论集合;通过How Net对词汇进行相似度计算,运用集对分析计算用户评论的联系度,实现基于语义的评论信息聚类;最后通过实验得到相关结论。
The huge amount of information of online user reviews is of great significance to the semantic mining of user comment content. The existing researches are mainly from the mining of product attributes and the analysis of user’s emotions in the reviews, and lack of semantic analysis. This paper presents a method of analysis that enables semantic mining of user comments. Firstly, the network user’s comment is abstracted into the form of “entity-attribute-value”, and the comment information is sentenced and segmented to form the semantic feature set of the comment; the similarity calculation of the vocabulary is made by HowNet, and the set pair analysis Calculating the degree of association of user comments, and implementing clustering of comment information based on semantic information; finally, the relevant conclusions are obtained through experiments.