Application of Semantic and Structural Similarity for Schema Reuse in Conceptual Database Design

摘要

Conceptual database design, aided by semantic and structural similarity, is based upon the reusability of concepts from existing database models. Objects selected from the existing database models are the foundation for a new database model. An initial version of the new model is extended with new objects and then it becomes a part of a library of models, aimed to help in the design of future models. A database, called multi-meta base, contains a description of the existing models. The design process is divided in two basic steps: the process of learning and the process of new model design. During the learning, an existing database model is included into the multi-meta base. A comparative analysis of the new model in respect to the existing database schemes is performed in order to detect any similarity between the objects. According to the detected similarity, the semantic connections among the objects are established. At the design of a new model, a selection of existing concepts is performed. The model is enhanced with new concepts, the consistency is tested and if necessary, a restructuring is done. Eventually, a semantic and a structural analysis of the new concepts in respect to the existing knowledge are performed.