Md. Kamrul Hasan et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5921-5928 A community-driven approach for missing background knowledge in Semantic matching Md. Kamrul Hasan Assistant Professor, Computer and Communication Engineering Department, Patuakhali Science and Technology University, Dumki, Patuakhali-8600, Bangladesh. Email: kamrul@pstu.ac.bd Quazi Delwar Hossain Assistant Professor, Electrical and Electronics Engineering Department, Chittagong University of Engineering and Technology, Bangladesh, Email: quazi@cuet.ac.bd A.H.M. Amimul Ahsan Faculty of Electrucal Engineering, UNIVERSITI TEKNOLOGI MALAYSIA, Malaysia E-mail: ahsan_utm@yahoo.com Mohammad Jamal Hossain Assistant Professor, Computer Science and Information Technology Department, Patuakhali Science and Technology University, Dumki, Patuakhali-8600, Bangladesh. Email: jamalpstu07@yahoo.com and Hasan Mahmud Department of Computer Science and Information Technology, Islamic University of Technology (IUT), Bangladesh, Email: hasan@iut-dhaka.edu Abstract Semantic heterogeneity is a key problem for information integration. An extensive work has been done to tackle this problem, and semantic matching is one of the proposed solutions. One challenge for matching algorithms is lack of background knowledge. This proposal presents an approach to deal with such drawback in semantic matching algorithms, which continuously improves quality results by involving community of users. Keyboards: Semantic Heterogeneity Problem, Ontology, Background Knowledge, Semantic Matching. I. Introduction Semantic heterogeneity problems are raised since organizations have local models and standards to represent knowledge, all of these evolving over time. Given the proliferation of information systems, and the growing specialization in key areas (e.g., bioinformatics, e-government), there exists a need to collaborate in order to build shared information systems, which benefit from the knowledge that can be obtained by combining information from different sources. This collaborative process poses the need of aligning different knowledge representations [2]. A lot of work has been done in building ontology matching solutions to automatically solve the semantic heterogeneity problems, see [1, 12] for surveys. Evaluation schemes on real schema-like structures, such as Google or Yahoo directories, show that there is still work to be done in order to improve quality of the matching solutions [8]. Semantic matching is a promising solution for ontology matching; unfortunately one drawback of both semantic and non-semantic approaches is the lack of domain specific background knowledge [9]. Some approaches try to discover this missing background knowledge automatically [9], and others rely on users to discover it, e.g., by asking them questions or showing them examples [14]. There is a growing trend in involving communities of users when trying to deal with matching inaccuracies, relying in these users to complement matching algorithms when these fail to find meaningful relations [3]. This proposal aims at presenting an approach ISSN: 0975-5462 5921