Query-Driven Conceptual Browsing: A Semi-Automated Approach for Building and Exploring Concepts on the Web

The presence of communities, which are groups of highly cross referenced pages together representing a single concept, is a striking feature of the World Wide Web. Quite often a group of communities, each topically coherent within itself, may be related through a common concept manifested in each of them. Motivated by this observation, we present a method for query-driven conceptual browsing for exploring concepts on the Web starting from a userspecified query. We show how this idea is related to prior work on learning concept maps and on Web Mining, and discuss the application of conceptual browsing for user-driven exploration and discovery of new concepts on the Web.