Exercise 4.8 Consider again the R dataset Boston already presented in Example 3.3 and Exercise 3.6 and study the relationship between the variable “House median value” (MEDV) and the variable “number of rooms” (RM). Estimate an a-spatial model using the OLS criterion, then estimate a Spatial Lag Model with the Maximum Likelihood procedure, finally estimate a Geographically Weighted Regression model using a Gaussian kernel and the bandwidth specified using the cross-validation method. Compare the results obtained with the various methods. In particular, compare the estimated regression coefficients obtained with OLS and SLM with the median value of the GWR coefficients. Compute Moran’s I statistics for the residuals of the GWR model.