This paper describes an experimental search engine based on a cluster map of user search terms. User search terms were collected from log files and e-mail reference questions. Clusters were produced using the SAS Average Linkage Hierarchical Cluster Procedure and four distance measures. A concept map was constructed using common elements from the clusters produced using the four distance measures. An advanced smart search based on the concept map can be called up when users need assistance broadening or narrowing their search for a web site cataloging links to water quality information.
|Library Philosophy and Practice
|Published - 2005
ASJC Scopus subject areas
- Library and Information Sciences