Web Relevant Term Suggestion Using Log-based and Text-based Approaches

Abstract

The Study attempts to integrate log-based and text-based methods ot obtain appropriate relevant terms for web search. The merits and limitations of using these two types of resources for term suggestion are also discussed. Further, the study proposes to cluster the relevant terms extracted into certain concept clusters hierarchically, which allows users to browse the terms in a more intuitive and meaningful way. The integrated model will enhance effectiveness of current web information retrieval systems and benefit further research.