Bayesian Estimation For The Spatial Autoregressive Conditional Heteroscedasticity Model

In this paper,the spatial autoregressive conditional heteroscedasticity(SPARCH)model with explanatory variables is introduced,which can be used to fit spatial data with heteroscedasticity properties.The maximum likelihood and Bayesian method are considered to estimate parameters in this model.These two methods are also evaluated and compared by the Monte Carlo simulations.It is showed that these two methods are reliable,and the Bayesian method performs better.Finally,we apply the model to fit an agricultural data.The result shows that compared to other spatial models,the SPARCH model with explanatory variables performs better in fitting some empirical data.