Schema integration algorithms using data mining and an object oriented model

The proliferation of database applications has highlighted the problem of integration of existing databases into an integrated system as a major challenge in meeting enterprise information requirements. The earlier proposed integration techniques concentrate on automating the integration process leaving the responsibility of appropriately establishing the assertions that provide correspondences between component databases on the database administrator (DBA). This is a bottleneck for schema integration tools. This thesis is an attempt to automate the generation of assertions, which can greatly improve the performance of the complete schema integration process. We adopt a new approach, based on data mining techniques to automate the assertion generation, thus reducing the burden on the DBA. One other problem in schema integration is model heterogeneity which we address by using an object oriented model for the schema integration phase.