An Ant Colony Optimization Classifier System for Bacterial Growth

Motivation. Ant colony optimization is one of the most recent nature-inspired metaheuristics. The algorithm mimics cooperative foraging behavior of real life ants, has already exhibited superior performance in solving combinatorial optimization problems. In this work, we have explored the searching capabilities of this metaheuristic for learning classification rules in bacterial growth/no growth data pertaining to pathogenic Escherichia coli R31. Method. The algorithm iteratively discovers a set of classification rules for a given dataset. At any iteration level, each one of the software ants develops trial rules and a rule with highest value of quality measure is denoted as a discovered rule, which represents information extracted from the training set. The cases correctly covered by the discovered rule are removed from the training dataset, and another iteration is started. Guided by the modified pheromone matrix, the agents build improved rules and the process is repeated for as many iterations as necessary to find rules covering almost all cases in the training set. Results. The capability of the ACO algorithm is gauged by considering two real world datasets. The performance of ACO algorithm is compared with the performance of tree based C4.5 algorithm with respect to the predictive accuracy and the simplicity of discovered rules. In both these performance indices ACO algorithm compares very well with C 4.5. Conclusions. The results obtained on two real life data sets indicate that the algorithm is competitive and can be considered a very useful tool for knowledge discovery in a given database