一种基于OCC模型的文本情感挖掘方法

观点挖掘（或情感分析）作为面向网络社会媒体分析挖掘领域的一个核心研究课题，具有重要的研究意义和应用价值。针对传统观点挖掘方法存在的不足和局限性，本文设计并实现了一种基于OCC情感模型的观点挖掘方法。该方法首先采用统计方法，利用WordNet词典、句法依存关系及少量标注数据，自动构建情感维度词典；其次，对所构建的情感维度词典进行求精，通过语义、情感倾向的不一致性处理和非情感词的过滤，得到高质量的情感维度词典；最后，基于所得到的情感维度词典，结合OCC模型中情感维度值与情感类型的对应关系，生成6种主要的情感类型。实验方法表明，此方法在使用灵活性、可解释性和有效性上具有明显的优势。

OCC-model-based text-emotion mining method

Opinion mining, also called sentiment analysis, as one of the core research areas in the network-oriented social media analysis and mining domain, has important practical and research significance. Due to the weaknesses and limitations of traditional opinion mining methods, in this study, we designe and implemente an OCC emotion model-based opinion mining method for extracting emotion types from text. First, we adopte a statistical method to construct an emotion dictionary, based on candidate sets collected by the WordNet dictionary, as well as several syntactic dependent relationships and a small amount of annotated data. Next, we refine the constructed emotion-dimension dictionary to improve its quality by filtering out non-emotional words as well as emotional words that have conflicting syntactic or orientation. Lastly, we generate six main emotion types based on the obtained emotion-dimension dictionary combined with the corresponding relations between emotional dimensions and the different emotion types identified by the OCC model. Experimental results show that the proposed method has obvious advantages with respect to flexibility of usage, interpretability, and effectiveness.