Unsupervised naive Bayes for data clustering with mixtures of truncated exponentials

Abstract

In this paper we propose a naive Bayes model for unsupervised data clustering, where the class variable is hidden. The feature variables can be discrete or continuous, as the conditional distributions are represented as mixtures of truncated exponentials (MTEs). The number of classes is determined using the data augmentation algorithm. The proposed model is compared with the conditional Gaussian model for some real world and synthetic databases.