图书情报工作 ›› 2016, Vol. 60 ›› Issue (16): 121-127,136.DOI: 10.13266/j.issn.0252-3116.2016.16.015

• 知识组织 • 上一篇    下一篇

基于特征本体的微博产品评论情感分析

唐晓波, 兰玉婷   

  1. 武汉大学信息管理学院 武汉 430072
  • 收稿日期:2016-05-05 修回日期:2016-08-02 出版日期:2016-08-20 发布日期:2016-08-20
  • 作者简介:唐晓波(IRCID:0000-0001-5885-4509),教授,博士生导师;兰玉婷(ORCID:0000-0003-3664-9381),硕士研究生,E-mail:453198036@qq.com。
  • 基金资助:

    本文系国家自然科学基金项目“社会化媒体集成检索与语义分析方法研究”(项目编号:71273194)研究成果之一。

Sentiment Analysis of Microblog Product Reviews Based on Feature Ontology

Tang Xiaobo, Lan Yuting   

  1. School of Information management, Wuhan University, Wuhan 430072
  • Received:2016-05-05 Revised:2016-08-02 Online:2016-08-20 Published:2016-08-20

摘要:

[目的/意义] 微博平台产品评论的特征级情感分析问题具有其特殊性,为了对特征分类,解决隐式特征的识别问题,并分析特征情感,提出一种基于特征本体的产品评论情感分析方法。[方法/过程] 该方法利用构建的特征本体对特征词分类,通过计算情感词与特征的搭配权重来识别隐式特征,并构建领域情感词典和微博表情符号词典,计算微博产品评论的特征情感极性和强度。[结果/结论] 构建方法模型,通过采集微博评论数据设计实验,验证了提出方法的有效性。

关键词: 特征, 本体, 产品评论, 情感分析, 微博

Abstract:

[Purpose/significance] Feature-level sentiment analysis on product reviews on the microblog platform has its particularity. To group features into categories, solve the recognition problem of implicit features and analyze features' sentiment, this paper proposes a sentiment analysis method of product reviews based on the feature-ontology. [Method/process] Feature-ontology is applied to grouping features into categories. Implicit features are identified by calculating collocation weights between emotional words and features. Emotion dictionary of a specialized field and microblog emoticon dictionary are constructed to calculate emotional intensity and polarity of features. [Result/conclusion] The methodological model is constructed and its effectiveness is verifies by collecting microblog reviews and designing the experiment.

Key words: feature, ontology, product reviews, sentiment analysis, micorblog

中图分类号: