中文领域情感词典构建研究  

The Construction of a Domain Sentiment Lexicon Based on Chinese Text

作　　者：王召义[1,2] 陈应红 周海燕 孙婷婷 Wang Zhaoyi;Chen Yinghong;Zhou Haiyan;Sun Tingting(Department of Economy and Trade,Anhui Business College,Wuhu Anhui 241002;Market Research and Data Service Center,Anhui Business College,Wuhu Anhui 241002)

机构地区：[1]安徽商贸职业技术学院经济贸易系,安徽芜湖241002 [2]安徽商贸职业技术学院市场调查与数据服务中心,安徽芜湖241002

出　　处：《情报探索》2020年第11期48-56,共9页Information Research

基　　金：2019年安徽省高校自然科学研究重点项目“电商在线评论情感指数分析与应用研究”(项目编号:KJ2019A007);2018年安徽省高校自然科学研究重点项目“基于Python编程的电商在线评论文本情感分析与应用研究”(项目编号:KJ2018A0721)成果之一。

摘　　要：[目的/意义]旨在为解决某些特定领域的情感分析任务,提高情感分析的精度提供参考。[方法/过程]采用逆向思维,先从领域情感词典结构设计方面入手,再以“实体-属性-情感词”的多元词对为主线,逐步解决领域情感词典结构要求,从而形成一种新的中文领域情感词典构建方法。[结果/结论]实验证明,该词典不仅能够准确判断情感词的情感极性,还能有效的对中文文本进行情感分类。[Purpose/significance]The paper is to provide references for solving some specific domain sentiment analysis task and improving accuracy of sentiment analysis.[Method/process]The paper adopts reverse thinking starts with the structural design of domain sentiment lexicon and then gradually solves the structural requirements of domain sentiment lexicon by taking the multi-word pair of“entity-attribute-sentiment word”as the main line thus forms a new method of constructing domain sentiment lexicon in Chinese.[Result/conclusion]Experiments show that the lexicon can not only accurately judge the sentiment polarity of sentiment words but also effectively classify the emotion of Chinese texts.

关 键 词：情感分析 领域情感词典 情感词 属性 实体 

分 类 号：G254[文化科学—图书馆学]