Open renyurenyu opened 2 months ago
In fact, we need to predict every spatial unit(or pixel). But due to the spatial autocorrelation, nearby units will get very similar results. Thus, we can skip some neighbors to save computation time. So this parameter depends on the spatial autocorrelation.
For example, in the urban pollution and NPP example, the code "predRows<-seq(5,totalRow,5) predCols<-seq(5,totalCol,5)" and predRows<-seq(10,totalRow,10) predCols<-seq(10,totalCol,10). There is a difference here. When I calculated the cause and effect relationship, I found that the calculated cause and effect relationship diagram was very different with settings of "10" and "5". I would like to ask if there is any basis for reference in setting the parameters here?