Relative Ontology Modeling for Relation Inference

Abstract

This paper presents a modeling of a relative ontology for extracting relations among people from unstructured data and definition of logic which each class can possess. Moreover, this throws out interesting and considerable points about ontology theory especially case of class. Three reasons why the 'relative' classes are hard to precisely be defined by using previous ontology theory are first described. Afterward, we construct basic class hierarchy of relative ontology through extracting necessary concepts from WordNet and checking consistency. In addition, logics of each class are defined. The ontology can be used as fundamental data for extracting relation among people by analyzing diverse web documents. Moreover, it can be applicable to diverse social information mining through extension.