What the query told the link: the integration of hypertext and information retrieval

摘要: Traditionally hypertexts have been limited in size by the manual effort required to create hypertext links. In addition, large hyper–linked collections may overwhelm users with range of possible links from any node, only a fraction which be appropriate for given user at time. This work explores automatic methods link construction based on feedback collected during browsing. A fulltext search engine mediates linking process. Query terms that distinguish well among documents database become candidate anchors; are mediated passage–based relevance queries. The newspaper metaphor is used organize retrieval results. VOIR, software prototype implements these algorithms has browse 74,500 node (250MB) articles. An experiment conducted test relative effectiveness dynamic and user–specified Experimental results suggest link– queries more effective than user– specified retrieving relevant information. paper concludes discussion extensions algorithms.