Automated Generalization of Fuzzy Concept Hierarchies for Attribute-Oriented Induction Purposes

This paper presents an approach to produce generalization candidates for a concept hierarchy without the necessity of being an expert in the domain to be generalized and its application to the summarization of large descriptive data sets. Through the use of ontologies, a set of terms can be cally generalized into the next abstract level in a concept hierarchy. A new approach for the extraction of y abstract generalizations of provided terms from ontologies, and for unsupervised construction of a fuzzy concept hierarchy for attributeoriented purposes will be presented. The algorithm presented is a proofofconcept type and it will be followed by future research. KEY WORDS Attributeoriented induction, data mining, ontology, fuzzy concept hierarchy, WordNet

• Type: Conference
• Year: 2006
• Where: DBA
• Authors: Jacob Dolan, Rafal A. Angryk
• Keywords: Concept Hierarchy, Database, DBA 2006, Descriptive Data Sets, Fuzzy Concept Hierarchy