基于点击模型的查询扩展方法
首发时间:2017-09-12
摘要:文章引入搜索用户点击模型,提出了一种基于用户点击模型的查询扩展方法,该方法首先从用户查询日志挖掘出所有关联的查询句对,并以此为初步查询扩展候选,然后将用户查询的一系列点击行为抽象为表述该查询的语义特征,并使用这些点击行为语义特征构建用户查询的点击向量,最后通过提出的一种综合相似度测度来衡量查询扩展的置信度。实验结果表明,该方法相比于传统的查询扩展方法可以有效改善搜索引擎相关性结果的语义匹配问题。
关键词: 查询扩展;信息检索;点击模型;日志挖掘
For information in English, please click here
query expansion method based on click model
Abstract:The article introduces the click model of search user,proposing a query expansion method based on user click model,firstly,the method extracts all the associated query sentence pairs from the user query log,and as a preliminary query extension candidate,then, a series of click behaviors of user queries are abstracted to express the semantic features of the query,and use these click action semantic features to construct the click vector of user queries,finally, we propose a synthetic similarity measure to weigh the confidence of query expansion.Experimental results show that this method can effectively improve the semantic matching problem of search engine correlation results compared to traditional query expansion methods.
Keywords: query expansion information retrie v a l click model log mining
基金:
引用
No.4741245802070150****
同行评议
共计0人参与
勘误表
基于点击模型的查询扩展方法
评论
全部评论0/1000