Knowledge bases: integration, verification, and partitioning

The literature is relatively sparse regarding the application of techniques such as optimization, graph theory and clustering analysis to knowledge base partitioning. A key question that needs to be addressed to apply cluster analysis to knowledge bases is the proper definition of the distance measure used to determine the similarity between rules in the knowledge base. A key question in applying graph analysis to knowledge base partitioning is the character of the connectivity between the rules, on one level, and between the partitions, on a higher level. Another issue involves the integration of the knowledge partitioning techniques into a broader methodology for knowledge base design.