分面搜索研究综述

Review on the Study of Faceted Search

传统的信息检索研究重点是基于精确查询的检索技术,每次搜索会返回大量的结果,这些结果可能相关也可能相关度不高,成功的搜索结果排序算法通常会返回前K个最相关的搜索结果来满足绝大部分查询。然而大多数的搜索任务在本质上是探索性的、不精确的,用户需要探索信息库,发现概念之间的关系,以引导方式缩小选择范围。一种探索式访问范式,分面搜索应运而生,它的出现解决了传统搜索方法的不足,为交互式信息检索奠定了基础。文中介绍了分面搜索的定义,并围绕分面搜索的关键技术、研究热点、最新进展和应用前景进行了探讨。

For the traditional information retrieval, researchers focus on the precise inquiry technology, by which a large number of results, which may not be highly correlated, are returned for each search. The successful search results sorting algorithm usually returns the top K relevant results to meet most of the inquiry. However, most real search tasks are exploratory and inaccurate, and the user needs to explore the information space, discover the relationship between concepts, and narrow down the scope of choice in a guided way. So a exploring access para- digm called faceted search has emerged to satisfy this need. Its appearance overcomes theshortcomings of the tradi- tional methods and lays a foundation for interactive information retrieval. This paper presents some definitions of fac- eted search and makes a study around the key technology, hot points, the latest progress and application prospect of faceted search.