XSAR: XML BASED SEARCH AGENT FORINFORMATION RETRIEVAL
Akhilesh Bajaj, Hideaki Tanabe, Chao-Chin Wang · 2002
The current form of information on the world wide web (WWW) is mainly in the form of unstructured information on scattered web sites. The usual method of accessing this information is using keyword-based or directory-based search engines. Recently, an increasing number of organizations, bodies and associations are adopting XML (extensible markup language) document type definitions (DTD) with a view towards putting structured information on the WWW. As information becomes more structured, we anticipate the spread of large XML repositories that will be accessible on the WWW. To help search these repositories, we present a novel search tool: XSAR (XML Based Search Agent For Information Retrieval), to access these repositories. We describe the architecture and functionality of XSAR, analyze its performance along four metrics and compare it to alternate existing search mechanisms. This work demonstrates the feasibility of agent based search mechanisms on large information repositories.