A Bayesian approach to learning mixtures of nonparametric components
By: Yilei Zhang , Yun Wei , Aritra Guha and more
Mixture models are widely used in modeling heterogeneous data populations. A standard approach of mixture modeling is to assume that the mixture component takes a parametric kernel form, while the flexibility of the model can be obtained by using a large or possibly unbounded number of such parametric kernels. In many applications, making parametric assumptions on the latent subpopulation distributions may be unrealistic, which motivates the need for nonparametric modeling of the mixture components themselves. In this paper we study finite mixtures with nonparametric mixture components, using a Bayesian nonparametric modeling approach. In particular, it is assumed that the data population is generated according to a finite mixture of latent component distributions, where each component is endowed with a Bayesian nonparametric prior such as the Dirichlet process mixture. We present conditions under which the individual mixture component's distributions can be identified, and establish posterior contraction behavior for the data population's density, as well as densities of the latent mixture components. We develop an efficient MCMC algorithm for posterior inference and demonstrate via simulation studies and real-world data illustrations that it is possible to efficiently learn complex distributions for the latent subpopulations. In theory, the posterior contraction rate of the component densities is nearly polynomial, which is a significant improvement over the logarithm convergence rate of estimating mixing measures via deconvolution.
Similar Papers
Bayesian nonparametric modeling of mixed-type bounded data
Methodology
Helps understand mixed health data better.
From Partial Exchangeability to Predictive Probability: A Bayesian Perspective on Classification
Methodology
Helps computers guess better with less data.
Nonparametric Bayesian Calibration of Computer Models
Methodology
Improves computer predictions for science and engineering.