A Hierarchical Clustering Approach for the Creation of a Simple Semantic Web Application

Abstract

The goal of the Semantic Web is to develop enabling standards and technologies designed to receive more exact results when searching for information, and to help machines understand more information on the Web so that they can support richer discovery, data integration and navigation. This can be achieved if there is a common vocabulary for a set of domains. Information is published using standard vocabulary. This study explores the processes of creating a taxonomy for a set of journal articles using hierarchical clustering algorithm. 100 journal articles that cut across different fields were downloaded from the internet. These served as sample data. These journal articles were serialized, stemmed and tokenized. Term frequency was calculated for each journal article.  Some representative terms were selected from each journal article and similarity matrix was generated for the entire journal articles. Complete hierarchical clustering was used to create a cluster of the articles. JavaTree view program was used to view the dendrogram of the cluster. It was observed that the articles cluster around their subject, subject area, field of study, area of application, journal type, author, place of case study. This demonstrated that journal articles have properties on a taxonomy, could be created as a basis for a semantic web.

Keywords: Semantic web, clustering, taxonomy, similarity, document collection.