结合依存关联分析和规则统计分析的情感词库构建方法

结合依存关联分析和规则统计分析的情感词库构建方法 The Study of Construction for Emotion Thesaurus Based on Dependency Parsing Combined with Rules and Statistics Methods 李勇敢 LI Yonggan 1 2 first-author 周学广 ZHOU Xueguang 3 孙艳 SUN Yan 3 张焕国 ZHANG Huanguo 1 武汉大学计算机学院 武汉大学计算机学院 School of Computer,Wuhan University School of Computer,Wuhan University 河南质量工程职业学院计算机系 河南质量工程职业学院计算机系 Department of Computer,Henan Vocational School of Quality Engineering Department of Computer,Henan Vocational School of Quality Engineering 海军工程大学信息安全系 海军工程大学信息安全系 Department of Information Security,Naval University of Engineering Department of Information Security,Naval University of Engineering 为了夯实自媒体时代的信息内容安全基础,本文把依存分析方法和传统的规则统计方法结合起来,构建了一个新型情感词库———钮库,与国内同行的情感新词识别公开评测结果比较,使用钮库辅助的本文方法表现优良,主要技术指标:微平均的准确率、召回率和F值,都在作者原有方法基础上提高了44%以上. To tamp the foundation of information content security for self-medium,this paper built a new emotion thesaurus,abbreviated as NUE Thesaurus,which combined with dependency parsing method and traditional rules and statistical methods.To compare with the results of an open evaluation of domestic counterparts,the proposed method,aided with NUE Thesaurus,presents excellent.The main technical indicators:micro-average precision,recall and F m easure increase by 44%compaled with rules and statistie method,which illustrate the effectiveness of the method. 中文信息处理 信息内容安全 依存分析 钮库 新词识别方法 Chinese information processing information content security dependency parsing NUE Thesaurus new words identification 10.14188/j.1671-8836.2013.05.011 TP391.1 国家自然科学基金（91018008,60970115）;全军军事学研究生课题（2011JY-434）的资助项目 2013-05-01 2021-04-01 5 59