Clustering RDF data using K-medoids

Abstract

Semantic web is a knowledge graph formed around semantic languages to enable computers and software to understand contents on the web. The content is explicitly annotated with semantic metadata using Resource Description Framework (RDF) language. However, the main issue is how to efficiently retrieve the RDF data taking into account a wide variety semantic and syntax nature and large-scale of such data. This paper aims to introduce a novel mechanism based on K-medoids algorithm for narrowing down the contents of the Web to clusters pertaining subset of information. We integrated sequence alignment algorithms with linguistic similarity measures to build a distance matrix which is used later in K-medoids clustering algorithm. The experimental outcomes showed a promised result for accuracy and quality of clustering.