Association rule mining using a bacterial colony algorithm

Bacterial colonies perform a cooperative distributed exploration of the environment. This paper describes bacterial colony networks and their skills to explore resources as a tool for mining association rules in databases. The proposed algorithm is designed to maintain diverse solutions to the problem at hand, and its performance is compared to other well-known bio-inspired algorithms, including a genetic and an immune algorithm (CLONALG) and, also, to Apriori over some benchmarks from the literature.