结合领域知识的中文句子评价对象抽取  

Integrate domain knowledge to extract opinion targets from Chinese sentences

作　　者：雷志城[1] 廖祥文[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350116

出　　处：《福州大学学报（自然科学版）》2013年第3期297-304,共8页Journal of Fuzhou University(Natural Science Edition)

基　　金：福建省自然科学基金资助项目(2010J05133);福建省科技创新平台计划资助项目(2009J1007)

摘　　要：针对评价对象存在领域相关性这一特点,在条件随机场模型中结合领域词词典特征进行中文句子评价对象的抽取,然后利用领域规则对抽取结果进行处理.针对COAE2011任务三标注语料的抽取实验结果表明,结合领域词词典和领域规则对于利用线性链、跳跃链和层叠条件随机场模型的中文句子评价对象抽取方法可以有效地提高抽取的精度,并抽取出更多的评价对象.In this paper we focus on the feature that opinion targets are domain related,and employ conditional random fields（CRFs） to extract the opinion targets incorporating domain dictionary,then we use domain rules in post processing.The method is evaluated on the labeled corpus of task 3 in COAE2011.The experiment results show that when employing the model of linear-chain,skip chain and cascaded CRFs to extract opinion targets from Chinese sentences,it is helpful to integrate both the domain dictionary and the rules.Our method can effectively improve the accuracy and extracts more objects simultaneously.

关 键 词：中文句子 评价对象抽取 领域知识 领域词词典 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]