摘要
近年来,微博在反映社情民意、丰富信息服务等方面发挥着越来越重要的作用.本文基于情感词典和依存文法,提出使用意见模型的方法表示微博文本,将微博中的句子形式化地表示为若干情感元素的组合,首先借助情感词典和修饰词典计算每个情感元素的情感倾向,进而通过加权求和法计算微博语句的总体情感倾向.实验表明该意见模型算法可以有效地判定微博文本情感倾向,以此为基础的微博情感分析系统能很好的支持校园管理的智能决策.
In recent years, micro blog play an increasingly important role in reflecting social conditions and public opinions. This paper proposes an opinions model to represent micro blog sentences based on emotional dictionary and dependency grammar, in which every sentence is represent as a group of opinion elements. We calculate the sentiment polarity score of every emotional element based on emotional dictionary and modifier dictionary, and then get a weighted summation sentiment evaluation for each micro blog sentence. The test show that the algorithm could effectively classify the micro blog sentiment tendencies and help to improve the performance of the intelligent decision system for campus management.
出处
《河北工业大学学报》
CAS
北大核心
2013年第6期24-29,共6页
Journal of Hebei University of Technology
基金
国家自然科学基金(61272511)
国家高技术研究发展计划(863)(2013AA01A212)
关键词
情感分析
微博
依存文法
智能决策系统
sentiment analysis
dependency grammar
micro blog
intelligent decision system
