Decision Rules Development Using Set of Generic Operations Approach

The main goal of presented research was to compile new approach for development learning models in a form of decision rule set. This approach devotes to using primary decision table as a primitive set of rules. Thus, each of learning cases is treated as a single classification rule. Next, a set of generic operations are applied to find the final, qualitative learning model. These generic operations are implemented in the RuleSEEKER system. During this research a few well-known algorithm for rule generation were compared with proposed solution. Obtained results are similar, sometimes even better and suggests that this method is a promising solution.