Abstract: To effectively utilize the product comment information of e-commerce platform and fully mine the feature level description of comments,a product-review feature mapping model based on aspect level emotion analysis was proposed.The text information and digital information in the product review were fully used by parsing comment text,extracting the entity feature-description candidate set and using the structured comment expression method to standard digital information.On this basis,the fine-grained emotion was excavated,the emotion dictionary and evaluation dictionary were built,and the aspect level emotion tendency was quantified to create the comment feature library.Then we build the product-review feature mapping model was constructed by combining comment feature library and the product composition performance tree.Finally,a domestic mobile phone was taken as an example to verify the feasibility and effectiveness of the proposed model.

Key words: structured comment expression, aspect-level emotion analysis, comment feature library, product-review feature mapping model

摘要： 为有效利用电商平台产品评论信息,充分挖掘评论的特征级描述,提出一种基于方面级情感分析的产品评论特征映射模型。首先,对评论文本进行句法分析,提取出实体特征描述候选集,使用结构化信息表达方法,并对数字信息进行标准化转换,实现对产品评论中文本信息和数字信息的充分利用。在此基础上进行细粒度的情感挖掘,构建情感词典和评价词典,对方面级情感倾向进行量化计算,用于创建评论特征库,结合产品组成性能树,构建产品评论特征映射模型。最后以某国产手机为对象进行实例研究,验证了所提出模型的可行性与有效性。

关键词: 结构化评论表达, 方面级情感分析, 评论特征库, 产品评论特征映射模型