查询意图研究综述  

Review of Research on Query Intent

作　　者：陆伟[1] 周红霞[2] 张晓娟[2] 

机构地区：[1]武汉大学信息资源研究中心 [2]武汉大学信息管理学院

出　　处：《中国图书馆学报》2013年第1期100-111,共12页Journal of Library Science in China

基　　金：教育部人文社科基地重大项目"面向细粒度的网络信息检索模型及框架构建研究"(项目编号:10JJD630014);国家自然科学基金面上项目"基于语言模型的通用实体检索建模及框架实现研究"(项目编号:71173164)的研究成果之一

摘　　要：近年来,学界对查询意图进行了广泛探讨,一些重要国际会议如SIGIR,WWW等越来越重视查询意图的相关研究,其研究点主要集中在给定分类体系下的查询意图识别,内容涉及查询意图类目体系构建、特征识别、分类方法以及数据集与评价方法。研究发现当前查询意图研究面临如下问题和挑战:缺乏权威的评测标准,各种分类方法在大规模查询集合上的性能还不确定,有效提取或者获得查询特征的方法仍然值得深入研究,查询意图分类体系的完备性和类别间独立不相关性尚不确定。Query intent has recently been deeply studied in academics, in particular when there is a given taxonomy, so more and more important international conferences such as SIGIR and WWW pay much attention to it. The focus of this research is the query intent recognition given the category system, which includes category building, feature identification, classification methods as well as datasets and evaluation methodology. The problems and challenges existing in query intent research are as follows： lack of authoritative evaluation indexes, the uncertainty of classifying performance on mass query sets, how to more effectively ex- tract or obtain query characteristics, the uncertainty of taxonomic completeness and the uncertainty of independent relationship between the classes.

关 键 词：查询意图 查询分类 查询特征 

分 类 号：G353.4[文化科学—情报学]