Implementation of Tuned Schema Merging Approach

• Main Authors: Nayyer Masood, Gul Jabee
• Format: Article
• Language: English
• Published: Mehran University of Engineering and Technology 2018-10-01
• Series: Mehran University Research Journal of Engineering and Technology
• Online Access: http://publications.muet.edu.pk/index.php/muetrj/article/view/558
• ISSN: 0254-7821; 2413-7219

Schema merging is a process of integrating multiple data sources into a GCS (Global Conceptual Schema). It is pivotal to various application domains, like data ware housing and multi-databases. Schema merging requires the identification of corresponding elements, which is done through schema matching process. In this process, corresponding elements across multiple data sources are identified after the comparison of these data sources with each other. In this way, for a given set of data sources and the correspondence between them, different possibilities for creating GCS can be achieved. In applications like multi-databases and data warehousing, new data sources keep joining in and GCS relations are usually expanded horizontally or vertically. Schema merging approaches usually expand GCS relations horizontally or vertically as new data sources join in. As a result of such expansions, an unbalanced GCS is created which either produces too much NULL values in response to global queries or a result of too many Joins causes poor query processing. In this paper, a novel approach, TuSMe (Tuned Schema Merging) techniqueis introduced to overcome the above mentioned issue via developing a balanced GCS, which will be able to control both vertical and horizontal expansion of GCS relations. The approach employs a weighting mechanism in which the weights are assigned to individual attributes of GCS. These weights reflect the connectedness of GCS attributes in accordance with the attributes of the principle data sources. Moreover, the overall strength of the GCS could be scrutinized by combining these weights. A prototype implementation of TuSMe shows significant improvement against other contemporary state-of-the-art approaches.