基于LM-NE的改进的相关反馈系统
首发时间:2011-11-29
摘要:在Internet中,相关反馈是信息检索中的一个重要的应用。实现相关反馈的一个方法是进行查询扩展(query expansion)。查询扩展的关键点是如何为反馈查询提供合适的扩展词。在本文中提到的相关反馈系统是基于语言模型(Language Model,简称LM)和命名实体(named entity,简称NE),简称为LM-NE系统。在本系统中,语言模型用来进行扩展词挑选,与主题相关的命名实体也会用助于反馈查询。实验结果证明采用语言模型和KL距离挑选命名实体的算法可又改进查询结果。
For information in English, please click here
Improved relevance feedback based on LM-NE system
Abstract:In this paper, relevance feedback is an important application in information retrieval on Internet. One method for relevance feedback is query expansion. The key point of query expansion is how to select the expanded words for feedback. In this paper, our RF system is designed based on language model and named entity, we call it LM-NE RF system. In this system, Language model is a model for term selecting. And named entity would help for feedback since these entities are related to the topics. The experiment result shows that the algorithm using language model and named entity based on KL-divergence could improve the result.(10 Points, Times New Roman)
Keywords: Relevance Feedback Language Model Named Entity
论文图表:
引用
No.4451128614526132****
同行评议
共计0人参与
勘误表
基于LM-NE的改进的相关反馈系统
评论
全部评论0/1000