Generating, integrating, and activating thesauri for concept-based document retrieval

Abstract

A blackboard-based document management system that uses a neural network spreading-activation algorithm which lets users traverse multiple thesauri is discussed. Guided by heuristics, the algorithm activates related terms in the thesauri and converges of the most pertinent concepts. The system provides two control modes: a browsing module and an activation module that determine the sequence of operations. With the browsing module, users have full control over which knowledge sources to browse and what terms to select. The system's query formation; the retrieving, ranking and selection of documents; and thesaurus activation are described.