Ontology-Based Searching Over Multiple Networked Data Sources

Abstract

In this paper we propose an ontology model to describe the domain of searching over multiple data sources (SND). The ultimate goal of the model is to bridge the heterogeneity and to solve the distribution of the multiple networked data sources. Based on the model a query reconstruction algorithm is proposed to translate the initial query into different local queries suitable for individual data sources. And the result processing module defines grammar rules to extract the useful data returned by the individual data source, consolidates the retrieved data into one unified result, and presents the results in the web browser. In the ontology model two constraint rules are defined to evaluate the performance of SND and experiment results show that the federated retrieval offers a reliable and efficient reproduction with the retrieved results from the independent data sources. Moreover, the figures indicate the time comparisons among multiple data sources and give the contribution of each data source by the returned searching results based on the same query set.