LTLAB中文微博情感分析评测报告-中国计算机学会中文信息技术专业.PDF
LTLAB 中文微博情感分析评测报告
1 1 1, †
周霄 周振宇 李芳
1. 上海交通大学电子信息与电气工程学院计算机科学与工程系,上海 200240
† 通信作者,E-mail: fli@
摘要 本文介绍了上海交通大学中德语言技术联合实验室(LTLAB )参加2012 年中文微博情感分析评测的方
法实现。在本届评测设立的3 个评测任务中,LTLAB 分别参加了任务1 (观点句识别)和任务3 (情感要素抽
取):对于任务1,参评系统使用了基于分类器的方案,特征抽取时考虑到了文本中的词性和句法特征;对于任
务3,参评系统结合了基于模板的抽取和基于分类器的抽取,考虑到了微博特有的话题信息以及词的统计信息。
评测结果表明,本文提出的方法在实践中是行之有效的。
关键词 微博; 情感分析; 观点句识别; 情感要素抽取
中图分类号 X123
LTLAB at Chinese Microblog Sentiment Analysis Track
1 1 1, †
ZHOU Xiao ,ZHOU Zhenyu , LI Fang
1. Department of Computer Science and Engineering, School of Electronic Information and Electrical Engineering,
Shanghai Jiao Tong University, Shanghai 200240
†Corresponding author, E-mail: fli@
Abstract This paper introduces the system that used by UDS-SJTU Joint Research Lab for Language Technology to
participate in the first Chinese Microblog Sentiment Analysis Track 。LTLAB attended 2 out of all 3 tasks, which are
task 1(identifying opinionated sentences) and task 3(extracting sentiment element) 。In task 1, system takes advantages
of POS tag and syntactic information, and uses a method based on classification to solve the problem. In task 3,
classification based method and pattern based method are combined to provide final result; Topic information and
statistical information of words are also considered. Evaluation result reveals the soundness of proposed system.
Key words Microblog; Sentiment analysis; Identifying opinionated sentence; Extracting sentiment element
1 引言
在本届