基于词语情感隶属度特征的情感极性分类.pdf

网络出版地址：/kcms/detail/11.2442.N1232.005.html 北京大学学报(自然科学版) Acta Scientiarum Naturalium Universitatis Pekinensis doi: 10.13209/j.0479-8023.2016.004 基于词语情感隶属度特征的情感极性分类 宋佳颖 黄旭 付国宏† 黑龙江大学计算机科学技术学院 , 哈尔滨 150080; † 通信作者 , E-mail: ghfu@ 摘要 在模糊集合论框架下探索基于词语情感隶属度的情感极性分类特征表示方法。首先以 TF-IDF 为权重 分别构建情感特征词语的正向、负向极性隶属度, 并以隶属度对数比作为分类特征值构建基于支持向量机的 情感极性分类系统。在产品评论、NLPCC2014 情感分类评测数据和 IMDB 英文影评等数据上的实验结果表 明, 基于情感隶属度特征的系统优于基于布尔、频度和词向量等特征表示的系统 , 从而验证了所提出的基于 情感隶属度特征表示的有效性。 关键词 情感极性分类 ; 模糊集合论 ; 隶属度 ; 支持向量机 中图分类号 TP391 Exploiting Lexical Sentiment Membership-Based Features to Polarity Classification † SONG Jiaying, HUANG Xu, FU Guohong School of Computer Science and Technology, Heilongjiang University, Harbin 150080; † Corresponding author, E-mail: ghfu@ Abstract A lexical sentiment membership based feature representation was presented for Chinese polarity classification under the framework of fuzzy set theory. To this end, the TF-IDF weighted words are first used to construct the corresponding positive and negative polarity membership for each feature word. Then, the log-ratio of each membership is computed. Finally, a support vector machines based polarity classifier is built with the membership log-ratios as its features. Furthermore, the classifier is evaluated over different datasets