Hypernym-Hyponym Relation Extraction from Indonesian Wikipedia Text

Made Nindyatama Nityasya, Rahmad Mahendra, Mirna Adriani

Abstract: This study aims to automatically acquire a collection of hypernym-hyponym word pairs in Bahasa Indonesia using pattern analysis approach. The Wikipedia text is utilized as the data source to obtain the hypernym-hyponym relations. The presented method to acquire relations is defined in systematic procedures, which are seed building, text pre-processing, sentence extraction, pattern extraction, and pattern matching. The number of extracted relations can be gradually increased using bootstrapping algorithm.