Large Scale Matching Issues and Advances

Introduction

Recently, we are witnessing an explosive growth of data in the business and scientific area. In fact, there are many databases and information sources available through the web covering different domains: semantic Web, deep Web, e-business, biology, digital libraries, etc. In such domains, the data generated are heterogeneous and voluminous e.g schemas with several thousand elements are common in e-business applications. Currently, the greatest challenge to take up is to perform the integration of such heterogeneous collections of data. Matching techniques are solutions to automatically find correspondences between these data in order to allow their integration in information systems. Matching has found considerable interest in both research and practice. In fact, matching is an operation that takes data as input (e.g XML schemas, ontologies, relational database schemas) and returns the semantic similarity values of their elements. However, matching these data at large scale represents a laborious process. The standard approach trying to match the complete input schemas will often lead to shading off performance. Various schema matching systems have been developed to solve the problem semi-automatically. Since schema matching is a semi-automatic task, efficient implementations are required to support interactive user feedback. In this context, scalable matching becomes a problem to be solved.

This chapter describes new research works of large scale schema and ontology matching. In the past years there has been quite an amount of research in the area of matching both for database schemas and more recently for ontologies. Several surveys (Rahm& Bernstein, 2002, Shvaiko & Euzenat, 2005) have been proposed covering many of the existing approaches. The survey proposed by (Rahm& Bernstein, 2002) is devoted to a classification of schema matching approaches and a comparative review of matching systems. The survey exposed by (Shvaiko & Euzenat, 2005) presents, as well, a new classification taking into account some novel schema/ontology matching approaches. A number of approaches and principles have been developed for matching small or medium data (schemas or ontologies). A major challenge that is still largely to be tackled is to scale up semantic matching in two ways: to a large number of data to be aligned or matched and to very large data. While the former is primarily addressed in the database area, the latter has been addressed by researchers in schema and ontology matching. Based on this observation, we provide a survey of work in the large scale area that differs from those proposed by (Rahm& Bernstein, 2002, Shvaiko & Euzenat, 2005). We provide in our study the main features of a large scale matching. We survey, then, the existing matching approaches at large scale called holistic and Pair-wise and we show how these approaches deal with scalability problem. We discuss the several related strategies and topics of optimization techniques, machine learning algorithms, statistical algorithms, etc. We describe the existing schema/ontology matching tools in the literature and compare them. This analysis of state of the art techniques allows us to make some conclusions and observations about the existing matching approaches and systems.

This chapter is organized as follows. Section 2 presents the motivation of large scale matching problem. Section 3 discusses the large scale matching approaches and presents a classification. In section 4, we describe the large scale matching tools. Section 5 reports some future directions and section 6 concludes this chapter.