Comparison of Graph-Based and Logic-Based Multi-Relational Data Mining Nikhil S. Ketkar, Lawrence B. Holder, and Diane J. Cook   View Paper (PDF)   Return to December 2005, Volume 7, Issue 2 Abstract We perform an experimental comparison of the graph-based multi-relational data mining system, Subdue, and the induc- tive logic programming system, CProgol, on the Mutagene- sis dataset and various artificially generated Bongard prob- lems. Experimental results indicate that Subdue can signif- icantly outperform CProgol while discovering structurally large multi-relational concepts. It is also observed that CProgol is better at learning semantically complicated con- cepts and it tends to use background knowledge more effec- tively than Subdue. An analysis of the results indicates that the differences in the performance of the systems are a result of the difference in the expressiveness of the logic-based and the graph-based representations. The ability of graph-based systems to learn structurally large concepts comes from the use of a weaker representation whose expressiveness is inter- mediate between propositional and first-order logic. The use of this weaker representation is advantageous while learn- ing structurally large concepts but it limits the learning of semantically complicated concepts and the utilization back- ground knowledge. ©2006 Association for Computing Machinery