计算机应用 ›› 2015, Vol. 35 ›› Issue (5): 1310-1313.DOI: 10.11772/j.issn.1001-9081.2015.05.1310

• 人工智能 • 上一篇    下一篇

基于情感角色模型的文本情感分类方法

胡杨1, 戴丹1, 刘骊1, 冯旭鹏2, 刘利军1, 黄青松1,3   

  1. 1. 昆明理工大学 信息工程与自动化学院, 昆明 650500;
    2. 昆明理工大学 教育技术与网络中心, 昆明 650500;
    3. 云南省计算机技术应用重点实验室, 昆明 650500
  • 收稿日期:2014-11-26 修回日期:2014-12-25 出版日期:2015-05-10 发布日期:2015-05-14
  • 通讯作者: 黄青松
  • 作者简介:胡杨(1991-),男,江苏南通人,硕士研究生,主要研究方向:机器学习、文本情感分析; 戴丹(1992-),女,江西永新人,硕士研究生,主要研究方向:机器学习、自然语言处理; 刘骊(1979-),女,重庆人,副教授,博士,主要研究方向:机器学习、嵌入式技术、计算机图形处理; 冯旭鹏(1986-),男,河南郑州人,助理实验师,硕士,主要研究方向:搜索引擎技术; 刘利军(1978-),男,河南辉县人,讲师,硕士,主要研究方向:医疗信息服务; 黄青松(1962-),男,湖南长沙人,教授,主要研究方向:智能信息系统、信息检索.
  • 基金资助:

    国家自然科学基金资助项目(81360230);科技部科技型中小企业技术创新基金资助项目(13C26215305404).

Classification method of text sentiment based on emotion role model

HU Yang1, DAI Dan1, LIU Li1, FENG Xupeng2, LIU Lijun1, HUANG Qingsong1,3   

  1. 1. Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming Yunnan 650500, China;
    2. Educational technology and Network Center, Kunming University of Science and Technology, Kunming Yunnan 650500, China;
    3. Yunnan Key Laboratory of Computer Technology Applications, Kunming Yunnan 650500, China
  • Received:2014-11-26 Revised:2014-12-25 Online:2015-05-10 Published:2015-05-14

摘要:

针对传统情感分类方法因情感项指向不明引发的误判和隐藏观点遗漏等问题,提出一种基于评价对象情感角色模型的文本情感分类方法.该方法首先识别文本中的潜在评价对象,通过局部语义分析对潜在评价对象所在语句进行情感标注,确定潜在评价对象所在语句的正负极性,并定义其情感角色;然后,改进特征权值计算方法,将情感角色对应的倾向值融入模型特征空间中;最后,通过特征聚合对特征空间实现模型降维.实验结果表明,所提方法与提取强主观性情感项作为特征的情感分类方法相比,分类准确率约提高3.2%,可有效改善文本情感分类效果.

关键词: 文本情感分类, 向量空间模型, 局部语义分析, 情感角色, 特征聚合

Abstract:

In order to solve the problem of misjudgment which due to emotion point to an unknown and missing hidden view in traditional emotion classification method, a text sentiment classification method based on emotional role modeling was proposed. The method firstly identified evaluation objects in the text, and it used the measure based on local semantic analysis to tag the sentence emotion which had potential evaluation object. Then it distinguished the positive and negative polarity of evaluation objects in this paper by defining its emotional role. And it let the tendency value of emotional role integrate into feature space to improve the feature weight computation method. Finally, it proposed the concept named "features converge" to reduce the dimension of model. The experimental results show that the proposed method can improve the effect and accuracy of 3.2% for text sentiment classification effectively compared with other approaches which tend to pick the strong subjective emotional items as features.

Key words: text sentiment classification, Vector Space Model (VSM), local semantic analysis, emotion role, feature converge

中图分类号: