Ontological approach for searching multi datasource web services

Abstract

This thesis addresses the issue of Web Service search. This thesis focuses on solving the problem of searching for Web Services which are associated with relatively similar datasources. It attempts to overcome the search limitations of the UDDI standard Web discovery protocol which is based on simplest keyword search and have no primitives to look into datasource associated w ith Web Services. This thesis developed a searching framework that take into account the Web Service w ith variety of ontologies and through adopting techniques like schema matching and ontology merging; the developed prototype can find relatively similar datasources. The prototype also adopted techniques based on Collaborative Filtering to infer more datasources that are relevant to the search request based on relatively similar matches from other datasources. The prototype represents an extension to the capability of Apache Axis. The prototype has been tested on sample of locally published Web Services that have datasources on books. The developed searching techniques prove to be more flexible than other frameworks including UDDI.