Customized product user comment opinion mining and comprehensive quantification method and system

The invention belongs to the field of text sentiment analysis, and particularly relates to a customized product user comment opinion mining and comprehensive quantification method and system. On the basis of user viewpoint extraction, a user evaluation unit classification algorithm is designed in combination with an LDA topic model, a word2vec word vector and a k-means clustering algorithm, and evaluation unit summarization and induction are achieved. Meanwhile, an emotional intensity calculation method is provided, and user evaluation is quantified according to categories in combination with mathematical statistics. According to the opinion mining method and system, rapid analysis, summarization and emotion quantification of mass text information are realized, so that a user comment opinion mining result is more intuitive. 本发明属于文本情感分析领域，具体为一种定制产品用户评论意见挖掘与综合量化方法及系统。本发明在用户观点抽取的基础上，结合LDA主题模型、word2vec词向量和k-means聚类算法设计了一种用户评价单元归类算法，实现评价单元总结和归纳。同时，提出一种情感强度计算方法，并结合数理统计量将用户评价按照类别量化。该意见挖掘方法和系统实现海量文本信息的快速分析、总结和情感量化，使用