论文部分内容阅读
网络评论中的信息特征及情感倾向是一种重要信息,文章针对有的中文产品评论信息挖掘存在的不足,提出了一种基于词汇共现性的产品特征聚类技术与细粒度情感分析技术。在产品评论中,同类的产品属性会有多种多样的表述方式,文章将产品评论中的产品特征进行归类,且与以往基于句子的整体情感分析不同,提出了针对产品特征的更细粒度情感分析技术,并且对没有相应情感倾向的属性词做出合理处理。
Information features and sentiment tendencies in online reviews are an important piece of information. According to the shortcomings of some existing Chinese product review information mining, this paper proposes a product feature clustering technique and fine-grained sentiment analysis technology based on word coexistence. In the product reviews, similar product attributes will have a variety of expressions, the article will be product reviews in the classification of product features, and with the previous sentence-based overall emotional analysis, the proposed product features for a more granular Emotional analysis technology, and do not have the appropriate emotional attribute attribute words to make reasonable treatment.