Extensions to the Ant-Miner Classification Rule Discovery Algorithm

摘要

Ant-Miner is an ant-based algorithm for the discovery of classification rules. This paper proposes four extensions to Ant-Miner: 1) we allow the use of a logical negation operator in the antecedents of constructed rules; 2) we use stubborn ants, an ACO-variation in which an ant is allowed to take into consideration its own personal past history; 3) we use multiple types of pheromone, one for each permitted rule class, i.e. an ant would first select the rule class and then deposit the corresponding type of pheromone; 4) we allow each ant to have its own value of the α and β parameters, which in a sense means that each ant has its own individual personality. Empirical results show improvements in the algorithm's performance in terms of the simplicity of the generated rule set, the number of trials, and the predictive accuracy.