I am trying to create interaction plots to visualize the interaction effects in a spatial error model. I created the attached charts for a linear regression model in R, but the packages I used to create these do not work for spatial regression models specifically (they are errorsarlm objects, not lm). Is there a way to create charts like this to visualize interactions in spatial models? Or a way to obtain the confidence intervals and prediction intervals to then plot manually?
A check of the internet found r-spatial.org, but I don't know what's available there. What stands out to me from your charts is that you likely should consider heteroscedasticity as well. When predicted-yi values differ, the corresponding sigma for Yi residuals generally should also.
- Badges
- Science topic