Mining Interesting Classification Rules: An Evolutionary Approach

Automated discovery of rules is, due to its applicability, one of the most fundamental and important method in Knowledge Discovery in Databases(KDD). It has been an active research area in the recent past. This paper presents a classification algorithm based on Evolutionary Approach(EA) that discovers interesting classification rules in the form If P Then D. A flexible encoding scheme, genetic operators and a suitable fitness function to measure the goodness of rules are proposed for effective evolution of rule sets. The proposed algorithm is validated on several datasets of UCI data set repository and the experimental results are presented to demonstrate the effectiveness of the proposed scheme for automated rule mining.