摘要
在分析维吾尔语词性规则和语法特征的基础上,以维吾尔语评论性语句为研究语料,提出了一种基于Bootstrapping算法的意见挖掘关系抽取方法。在每一次迭代过程中,根据改进的评分公式选取最优模式抽取主题词-意见词对;迭代结束后,对于主题-意见词对为空的评论语句,使用最近匹配算法抽取主题-意见词对;用并联模式和否定模式对抽取的主题-意见词对进行扩展和修正。关系抽取的最终目标是为每一个评论性语句建立一个或多个二元组<主题词,意见词>,并使主题词和意见词一一对应。实验结果表明了该方法在关系抽取上的有效性。
On the basis of analyzing the Uyghur part-of-speech rules and grammatical characteristics, a relation extraction met- hod of opinion mining based on Bootstrapping algorithm is proposed, which take Uyghur comment sentences as the research cor- pus. In each iteration process, the optimal patterns are selected to extract topic-opinion pairs according to the improved score for- mulas. After the iteration, for the comment sentences that topic-opinion pairs are empty, the nearest matching algorithm is used to extract topic-opinion pairs. Finally, paralleling model and negation model are introduced to expand and amend topic-opinion pairs. The ultimate goal of relation extraction is to establish one or more tuples ~topic, opinion~ for every comment sentence, and make the topic word correspond to the opinion word. Experimental results show the effectiveness of the proposed method in relation extraction.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第9期3260-3265,共6页
Computer Engineering and Design
基金
国家自然科学基金项目(61262064
60963017
61063026
61063043)
国家社科基金项目(10BTQ045
11XTQ007)