The application of ant colony optimization to the classification rule problem

The goal of data mining is to extract knowledge from data. The discovered knowledge plays an important role in decision making. Accordingly, the comprehensibility of the discovered knowledge is very critical. If the discovered knowledge is not comprehensible, it will not be useful for decision making, even leads to an incorrect decision. Data mining is an interdisciplinary research topic. There are various data mining tasks. This paper focuses on the solving for classification rule problem. Classification rule is the most common representation of the rule in data mining. It belongs to the supervised learning process which generates rules from training data set. The goal of the classification rule mining is the prediction of the predefined class. Considering the classification task of data mining, discovered knowledge is often expressed in the form of IF-THEN rules. Rafael S. Parpinelli etc. proposed the Ant-Miner algorithm. Based on ACO algorithm, Ant-Miner solved the classification rule problem. In its basic configuration, Ant-Miner shows good performance in many dataset. In this paper, an extension of Ant-Miner is proposed to incorporate the concept of parallel processing and grouping. Intercommunication via pheromone among ants is a critical part in ant colony optimization´s searching mechanism. Due to the algorithm design, Ant-Miner made a slight modification in this part which removes the parallel searching capability. Based on Ant-Miner, we propose an extension that modifies the algorithm design to incorporate parallel processing. The pheromone trail deposited by ants during searching affected each other. With the help of pheromone, ants can have better decision making while searching. For solving the classification rule problem, we design an algorithm with the concept of multi-level rule choosing mechanism in order to get more accuracy of rule induced. Furthermore, we provide a possible direction for researches toward the classification rule problem.