Case-Based Reasoning for Data Warehouse Schema Design

There have been several efforts made in the area of semi-automated techniques and tools for schema design. Most have focused on database schema design, and little work has been done on data warehouse schema design. This paper presents a new approach to data warehouse schema design based on case-based reasoning theory. Case-based reasoning is a problem solving paradigm that involves the development of a solution space to provide a basis for efficient reuse of proven pre-existing solutions. A detailed system design for the application of case-based reasoning to data warehouse schema design is examined in this paper. Two novel contributions of the paper include the matching of cases by data warehouse schema meta-data, and the use of a business context aware ontology to make intelligent suggestions on entities and attributes. These two approaches assist in improving the case matching and adaptation capability in the case-based reasoning system.