Abstract: First, the linear Bayes was used to estimate the parameters of spatial autoregressive model, and the superiorities of the linear Bayes estimator over two-step least square estimator were studied in terms of the mean square error matrix (MSEM) criterion. Then, the estimation of spatial autocorrelation coefficient was implemented by Metropolis algorithm. Finally, the superiority of the linear Bayes estimation and two-step least square estimation was compared by simulation experiments.

Key words: spatial autoregressive model, linear Bayes estimation, two-step least square estimation