Bayesian estimation of heteroscedastic skew-normal error regression model

In statistics‎, ‎errors are inherent in data and models‎, ‎particularly heteroscedasticity and skew-normal error structures‎. ‎These errors were simultaneously generated and infused into the data‎, ‎leading to uncertainty in parameter estimation‎. ‎The statistician uses statistical knowledge to elicit information and guide decision-making‎. ‎Both classical and Bayesian restricted Stein-rule least squares were compared when the data were contaminated with the aforementioned errors‎. ‎This study proposed an innovative Bayesian generalized restricted Stein-rule least squares method with heteroscedastic skew-normal errors‎, ‎which was ultimately found to be more efficient compared to non-Bayesian restricted Stein-rule least square estimators‎. ‎The study observed excellent performance of the Bayesian frameworks‎, ‎including the Bayes estimate and posterior mean‎, ‎in comparison to the classical restricted Stein-rule least squares estimators‎. ‎Therefore‎, ‎the study recommends Bayesian generalized restricted Stein-rule least squares to analysts and researchers who may encounter such errors in their data‎.

کلیدواژه ها:

Bayesian ، ‎Heteroscedasticity ، ‎Modeling and least squares ، ‎Skew-normal‎ ، ‎Simulation