Bayesian Inference on Parameters for Mixtures of α-Stable Distributions Based on Markov Chain Monte Carlo
Junli Liang · Journal of Xi'an University of Technology · 2012
To solve the difficult problem of non-Gaussian signal difficult to be described,this paper suggests a method of Bayesian inference on parameter for mixtures of α-stable distributions based on Markov Chain Monte Carlo.The hierarchical Bayesian graph model is constructed.Gibbs sampling algorithm is used to achieve the estimation of the mixing weights and allocation parameter z.The 4 parameter estimations in each distribution component are completed on the basis of Metropolis algorithm.The simulation results show that the method can accurately estimate the parameters of mixture of α-stable distributions,and it has good robustness and flexibility,whereby the method can be used to establish the model for non-Gaussian signal or data.