A novel framework for the composition of schema matchers

  • Authors:

  • Balazs Villanyi;Peter Martinek;Bela Szikora

  • Affiliations:

  • Department of Electronics Technology, Budapest University of Technology and Economy, Budapest, Hungary;Department of Electronics Technology, Budapest University of Technology and Economy, Budapest, Hungary;Department of Electronics Technology, Budapest University of Technology and Economy, Budapest, Hungary

  • Venue:
  • ICCOMP'10 Proceedings of the 14th WSEAS international conference on Computers: part of the 14th WSEAS CSCC multiconference - Volume I
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The variety of enterprise schemas has entailed a reasonable interest in those methods, which should resolve this divergence of business entities in integration scenarios. In the case of schema matching, the goal is to find means to identify semantically related concepts. This procedure contributes to the success of the aforementioned task. Several solutions have been presented with reasonable performance qualities and genius behind lying concepts. Our investigations showed that the performance is strongly dependent on the idiosyncrasies of the actual test scenario. The not optimal parameter settings of the algorithm could result in a complete failure on the algorithm's part. On the other hand if the parameters are set correctly, the performance should significantly improve. Hence the key to success is the optimization of working parameters in the available algorithms. So far, we have introduced a new approach called calibration, which is a pre-run optimization technique providing an optimal parameter set to the process. In this paper we focus on the establishment of a new framework, which also facilitates the optimal composition of the input methods. Just like in the case of the calibration, we endeavor to provide optimization for given scenarios. Other prerequisites, such as the extensibility, scalability and maintainability were also considered.