Research on Sentiment Analysis Model of Douban Book Short Reviews based on Machine Learning

The era of big data has produced a large amount text data. Due to the different sources and complexity generation process, heterogeneous structure. Combining several latest machine learning methods with hierarchical technology, algorithm for heterostructure can improve accuracy classification. Empirical analysis shows that our significant effect on improving classification accuracy. Based grounded theory, this paper analyzes 2000 "short comments" sample books "Douban Reading" website, explores motivation short comments corresponding emotional analysis. results show users' tendencies in Douban applications are diversified consistent, neutral play an important role. user's tendency represents evaluation attitude towards books. be grasped timely manner. Readers used understand recognition manner adjust their service strategies