BeeMiner: A novel artificial bee colony algorithm for classification rule discovery

Abstract

Artificial bee colony (ABC) is a new population-based algorithm that has shown promising results in the field of optimization. In this paper, we propose BeeMiner, a novel ABC algorithm for discovering classification rules. BeeMiner differs from the original ABC because it uses an information-theoretic heuristic function (IHF) to guide the bees to search across the most promising areas of the search space. We compare the performance of BeeMiner with those of J48, JRip, and PART on nine benchmark datasets from the UCI Machine Learning Repository. The results show that BeeMiner is competitive with J48, JRip, and PART in terms of the predictive accuracy.