Analysis of Japanese wikipedia category for constructing wikipedia ontology and semantic similarity measure

Abstract

Wikipedia is a free encyclopedia on the Internet that contains millions of articles. Wikipedia ontologies (e.g., YAGO2 and the Japanese Wikipedia ontology) have been constructed to utilize the semantic information in this encyclopedia; in these ontologies, the category information from each article is used to construct concept hierarchies. However, Wikipedia categories were originally designed for navigation, to help users find appropriate articles, and they are not equivalent to concept hierarchies in Wikipedia ontologies. In this study, we briefly review the definition of a category in Wikipedia and exhaustively investigate the category structure in the Japanese Wikipedia. We also discuss how category structure information can be utilized to construct an ontology and to calculate a semantic similarity measure.