Knowledge Based Data Cleaning for Data Warehouse Quality

This paper describes an approach for improvement the quality of data warehouse and operational databases with using knowledge. The benefit of this approach is three-folds. First, the incorporation of knowledge into data cleaning is successful to meet the user's demands and then the data cleaning can be expanded and modified. The knowledge that can be extracted automatically or manually is stored in repository in order to be used and validated among an appropriate process. Second, the propagation of cleaned data to their original sources in order to validate them by the user so the data cleaning can give valid values but incorrect. In addition, the mutual coherence of data is ensured. Third, the user interaction with data cleaning process is taken account in order to control it. The proposed approach is based in the idea that the quality of data will be assured at the sources and the target of data.