How do I perform cross validation for a raster layer? Not a GA layer.

Question

I have raster layer and I am trying to perform cross validation so that I can understand the variance on a cell by cell level. It seems that you can only perform cross validation on a GA layer, however, this is not ideal, because it does not use pixels of a fixed size but instead uses contours and therefore does not have the same predictions as my raster.

Thanks

MichaelBeuthe · Answer

Thanks Steve,
I will try an elaborate. Let's say that I theoretically have a raster layer with cells that are 100 sq. ft in size and I want to use cross validation to see the variance between an empirical data point and the cell that that data point is located in (maintaining the size of the cell is important). Is there a tool that will allow me to do this? It seems that the only way I can use cross validation is on a GA layer. The problem with this is the GA layer uses contours rather then cells and does not allow me to set a resolution or pixel size. So if attaining variance date for a cell of specific size is important, then the CV results from the GA layer would not be very useful.

Hopefully this makes the question clearer.

Thanks

SteveLynch · Answer

Not sure that I understand your need for raster output. The cross validation is based on the sample locations.

A GA layer can be converted to raster if you want a fixed cell size.

-Steve