Title :
Tuned schema merging (TuSMe)
Author :
Jabeen, Gul ; Masood, Nayyer
Author_Institution :
Dept. of Comput. Sci., Int. Islamic Univ., Islamabad, Pakistan
Abstract :
Schema merging involves integration of multiple data sources into a global conceptual schema (GCS). It plays a significant role in different application domains, like multidatabase systems. A prerequisite to schema merging is schema matching, in which source databases are compared with each other to identify corresponding elements. There are different possibilities to create a GCS given a set of data sources and the correspondences between them. The tuned schema merging (TuSMe) is an effort to develop a balanced GCS that will control its horizontal and vertical expansion. Too much horizontal or vertical expansion results an unbalanced GCS that will produce too much NULL values against the global queries or cause poor query processing due to too many Joins required. TuSMe assigns weights to individual attributes that reflect the connectedness of attributes to the attributes of the underlying data sources. Theses weights are used to determine the internal cohesiveness of the individual relations in the GCS that can be combined to measure the overall strength of the GCS. TuSMe guides the human or automated integrator to establish a balanced GCS using the weights assigned to individual relations.
Keywords :
distributed databases; merging; data source integration; global conceptual schema; horizontal expansion; multidatabase system; schema matching; tuned schema merging; vertical expansion; Context; Distributed databases; Humans; Merging; Semantics; Unified modeling language; Database integration; global conceptual schema; schema merging; vertical and horizontal expansion;
Conference_Titel :
Software, Knowledge Information, Industrial Management and Applications (SKIMA), 2011 5th International Conference on
Conference_Location :
Benevento
Print_ISBN :
978-1-4673-0247-0
DOI :
10.1109/SKIMA.2011.6090010