A particle swarm data miner

Abstract. This paper describes the implementation of Data Mining tasks using Particle Swarm Optimisers. The object of our research has been to apply such algorithms to classification rule discovery. Results, concerning accuracy and speed performance, were empirically compared with another evolutionary algorithm, namely a Genetic Algorithm and with J48- a Java implementation of C4.5. The data sets used for ex-perimental testing have already been widely used and proven reliable for testing other Data Mining algorithms. The obtained results seem to indicate that Particle Swarm Optimisers are competitive with other evo-lutionary techniques, and could come to be successfully applied to more demanding problem domains.