Fuzzy view-based semantic search

Abstract

This paper presents a fuzzy version of the semantic view-based search paradigm. Our framework contributes to previous work in two ways: First, the fuzzification introduces the notion of relevance to view-based search by enabling the ranking of search results. Second, the framework makes it possible to separate the end-user's views from content indexer's taxonomies or ontologies. In this way, search queries can be formulated and results organized using intuitive categories that are different from the semantically complicated indexing concepts. The fuzziness is the result of allowing more accurate weighted annotations and fuzzy mappings between search categories and annotation ontologies. A prototype implementation of the framework is presented and its application to a data set in a semantic eHealth portal discussed. © Springer-Verlag Berlin Heidelberg 2006.