7 A Quality-enabled Spatial Integration System
155
sector. In order to limit the number of transactions, we gather for each sector the
reformulated subqueries according to their sources. The decomposition of the result
will be carried out at the mediator level. If a data source does not support a space
operation, the latter is executed at the mediator level. Once the subqueries are executed, we obtain their output data. These data are fused thereafter to obtain a result
by sector.
7.8 Conclusions
Traditional mediation systems do not take into account quality information in the
data integration process. Most of the time they assume that data are of equal quality
by means of some data-cleaning mechanism for instance. In this chapter, we described a quality-mediation approach that we developed in the context of a GIS integration project. Besides the well-known semantic heterogeneities, we also had to
address a very difficult problem, that is assessing data quality of the mediated geographical system. One of the problems is the complexity of query rewriting. Incorporating quality information in the rewriting algorithm increases the complexity. For
instance, some quality criteria are contradictory: improving the deficit rate implies
degradation of the surplus or the violation rates. The trade-off is to deliver a rather
complete set of answers without affecting its consistency. We believe our approach
has three main advantages. First, it allows dynamic combination of mapping rules accordingly with the quality requirements, leading to quality rating of the result before
data extraction. Second, it takes into account other metadata such as the spatial cover
in the query rewriting process, leading to a significant reduction of useless operations
(join of information related to different areas). Finally, it ensures system extensibility: adding or removing a data source does not significantly affect the global schema,
and may result in some new mapping rules to be inserted in the system.
Acknowledgments
We would like to acknowledge the French RNTL research framework for the support
of the VirGIS project.
References
1. Amann B, Beeri C, Fundulaki I, Scholl M (2002) Querying XML Sources Using an
Ontology-based Mediator. In: Proc. of the 10th Int. Conf. on Cooperative Information
Systems. Irvine, California, USA, 429–448
2. Base de Donn´ ees Topographiques du Qu´ ebec (2001) felix.geog.mcgill.ca/
heeslib/bdtq 20000.html
3. Beeri C, Kanza Y, Safra E, Sagiv Y (2004) Object Fusion in Geographic Information
Systems. In: Proc. of the 30th Int. Conf. on Very Large Data Base. Toronto, Canada,
816–827
155
sector. In order to limit the number of transactions, we gather for each sector the
reformulated subqueries according to their sources. The decomposition of the result
will be carried out at the mediator level. If a data source does not support a space
operation, the latter is executed at the mediator level. Once the subqueries are executed, we obtain their output data. These data are fused thereafter to obtain a result
by sector.
7.8 Conclusions
Traditional mediation systems do not take into account quality information in the
data integration process. Most of the time they assume that data are of equal quality
by means of some data-cleaning mechanism for instance. In this chapter, we described a quality-mediation approach that we developed in the context of a GIS integration project. Besides the well-known semantic heterogeneities, we also had to
address a very difficult problem, that is assessing data quality of the mediated geographical system. One of the problems is the complexity of query rewriting. Incorporating quality information in the rewriting algorithm increases the complexity. For
instance, some quality criteria are contradictory: improving the deficit rate implies
degradation of the surplus or the violation rates. The trade-off is to deliver a rather
complete set of answers without affecting its consistency. We believe our approach
has three main advantages. First, it allows dynamic combination of mapping rules accordingly with the quality requirements, leading to quality rating of the result before
data extraction. Second, it takes into account other metadata such as the spatial cover
in the query rewriting process, leading to a significant reduction of useless operations
(join of information related to different areas). Finally, it ensures system extensibility: adding or removing a data source does not significantly affect the global schema,
and may result in some new mapping rules to be inserted in the system.
Acknowledgments
We would like to acknowledge the French RNTL research framework for the support
of the VirGIS project.
References
1. Amann B, Beeri C, Fundulaki I, Scholl M (2002) Querying XML Sources Using an
Ontology-based Mediator. In: Proc. of the 10th Int. Conf. on Cooperative Information
Systems. Irvine, California, USA, 429–448
2. Base de Donn´ ees Topographiques du Qu´ ebec (2001) felix.geog.mcgill.ca/
heeslib/bdtq 20000.html
3. Beeri C, Kanza Y, Safra E, Sagiv Y (2004) Object Fusion in Geographic Information
Systems. In: Proc. of the 30th Int. Conf. on Very Large Data Base. Toronto, Canada,
816–827