Using Fuzzy Ontologies to Extend Semantically Similar Data Mining

Eduardo L. G. EscovarCristiane A. YaguinumaMauro Biajiz

Association rule mining approaches traditionally generate rules based only on database contents, and focus on exact matches between items in transactions. In many applications, however, the utilization of some background knowledge, such as ontologies, can enhance the discovery process and generate semantically richer rules. Besides, fuzzy logic concepts can be applied on ontologies to quantify semantic similarity relations among data. In this context, we extended SSDM (Semantically Similar Data Miner) algorithm in order to obtain from a fuzzy ontology the semantic relations between items. As a consequence, the generated rules can be more understandable, improving the utility of the knowledge supplied by them.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: