Bayesian GLS Regression with application to LP3 Regional Skew Estimation

摘要

This paper develops a Bayesian approach to Generalized Least Squares regression allowing computation of the posterior distributions of the parameters and the model-error variance using a quasi-analytic analysis and Markov Chain Monte Carlo simulation. Bayesian GLS analysis is illustrated with two regional skew estimation examples. The examples demonstrate that regional skew models can be relatively precise with effective record lengths in excess of 50 years. The marginal posterior distribution of the model error variance and the mean and variance of the parameters can be computed directly thereby providing a simple but important extension of the regional GLS procedures popularized by Tasker and Stedinger (1989).