Multi Objective Optimization of classification rules using Cultural Algorithms

Classification rule mining is the most sought out by users since they represent highly comprehensible form of knowledge. The rules are evaluated based on objective and subjective metrics. The user must be able to specify the properties of the rules. The rules discovered must have some of these properties to render them useful. These properties may be conflicting. Hence discovery of rules with specific properties is a multi objective optimization problem. Cultural Algorithm (CA) which derives from social structures, and which incorporates evolutionary systems and agents, and uses five knowledge sources (KS's) for the evolution process better suits the need for solving multi objective optimization problem. In the current study a cultural algorithm for classification rule mining is proposed for multi objective optimization of rules.