Dirichlet Prior for Estimating Unknown Regression Error Heteroscedasticity

We propose a Bayesian procedure to estimate heteroscedastic variances of the regression error term ω, when the form of heteroscedasticity is unknown. The prior information on ω is based on a Dirichlet distribution, and in the Markov Chain Monte Carlo sampling, its proposal density parameters' information is elicited from the well-known Eicker-White Heteroscedasticity Consistent Variance-Covariance Matrix Estimator. We present a numerical example to show that our scheme works.