Random Subclasses Ensembles by Using 1-Nearest Neighbor Framework Amir Ahmad * College of Information Technology United Arab Emirates University Al Ain, UAE amirahmad@uaeu.ac.ae Hamza Abujabal Faculty of Computing and Information Technology King Abdulaziz University Rabigh, Saudi Arabia C. Aswani Kumar School of Information Technology and Engineering VIT University Vellore 632014, Tamilnadu, India Received 25 November 2015 Accepted 13 February 2017 Published 19 April 2017 A classi¯er ensemble is a combination of diverse and accurate classi¯ers. Generally, a classi¯er ensemble performs better than any single classi¯er in the ensemble. Naive Bayes classi¯ers are simple but popular classi¯ers for many applications. As it is di±cult to create diverse naive Bayes classi¯ers, naive Bayes ensembles are not very successful. In this paper, we propose Random Subclasses (RS) ensembles for Naive Bayes classi¯ers. In the proposed method, new subclasses for each class are created by using 1-Nearest Neighbor (1-NN) framework that uses randomly selected points from the training data. A classi¯er considers each subclass as a class of its own. As the method to create subclasses is random, diverse datasets are generated. Each classi¯er in an ensemble learns on one dataset from the pool of diverse datasets. Diverse training datasets ensure diverse classi¯ers in the ensemble. New subclasses create easy to learn decision boundaries that in turn create accurate naive Bayes classi¯ers. We developed two variants of RS, in the ¯rst variant RS(2), two subclasses per class were created whereas in the second variant RS(4), four subclasses per class were created. We studied the performance of these methods against other popular ensemble methods by using naive Bayes as the base classi¯er. RS(4) outperformed other popular ensemble methods. A detailed study was carried out to understand the behavior of RS ensembles. Keywords : Classi¯er ensembles; naive bayes; subclasses; clusters; bagging; adaBoost.M1. * Corresponding author. International Journal of Pattern Recognition and Arti¯cial Intelligence Vol. 31, No. 10 (2017) 1750031 (26 pages) # . c World Scienti¯c Publishing Company DOI: 10.1142/S0218001417500318 1750031-1 Int. J. Patt. Recogn. Artif. Intell. Downloaded from www.worldscientific.com by MCMASTER UNIVERSITY on 05/30/17. For personal use only.