基于特征聚类的网络评论信息挖掘研究

Research on Mining Network Comments based on Feature Clustering

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。

The characteristics of information and emotional tendencies of comments in the network is a kind of important information, the Chinese product reviews existing mining problems, proposed one kind based on the lexical co-occurrence technology products feature clustering and fine-grained affective analysis technique. In product reviews, product attributes similar will have a variety of ways of expression, the product features product reviews are classified, and the whole emotional sentence analysis based on different, proposed a more fine-grained emotion analysis technique for product features, and on the attribute words without a corresponding emotional tendency to make the reasonable processing.