My knowledge of this topic is rather limited. But essentially i have a linear model based upon a number of predictor variables which i am running to produce an estimate.
I need to account for spatial autocorrelation within my data.
Am i understanding this correctly in that i can run my linear model through a GEE, without specifying coordinates for my data points and account for spatial auto correlation by specifying the distribution of the 'unknown'(spatial) correlation?
I have attempted other methods which account for spatial autocorrelation by processing the data with e.g. XCOR YCOR variables but R crashes. My sample size is over 2000.