Data Clustering using Variational Learning of Finite Scaled Dirichlet Mixture Models

Authors:

Abstract
We introduce a variational Bayesian learning framework for finite Scaled Dirichlet mixture model. The variational inference estimates the parameters without the computational cost of conventional Bayesian methods. Furthermore, our model avoids including additional model selection criterion by automatically determining the optimal number of components as a part of the inference process. The proposed model is tested with real-life data clustering problems to demonstrate its effectiveness.

This publication has 31 references indexed in Scilit: