Classi(cid:12)cation with Cluster-based Bayesian Multi-Nets using Ant Colony Optimization

Khalid M. Salama,A. Freitas

Abstract:Bayesian Multi-net (BMN) classi(cid:12)ers consist of several local models, one for each data subset, to model asymmetric, more consistent dependency relationships among variables in each subset. This paper extends an earlier work of ours and proposes several contributions to the (cid:12)eld of clustering-based BMN classi-(cid:12)ers, using Ant Colony Optimization (ACO). First, we introduce a new medoid-based method for ACO-based clustering in the Ant-ClustB MB algorithm to learn BMNs. Both this algorithm and our previously introduced Ant-ClustB IB for instance-based clustering have their effectiveness empirically compared in the context of the \cluster-then-learn" approach, in which the ACO clustering step completes before learning the local BN classi(cid:12)ers. Second, we propose a novel \cluster-with-learn" approach, in which the ACO meta-heuristic performs the clustering and the BMN learning in a synergistic fashion. Third, we adopt the latter approach in two new ACO algorithms: ACO-ClustB IB , using the instance-based method, and ACO-ClustB MB , using the medoid-based method. Empirical results are obtained on 30 UCI datasets.