I am running the forest based classification and regression tool (Spatial Statistics) and have trained the model using explanatory rasters and feature layers. I am now trying to predict to a feature layer and attempting to input the matching explanatory rasters with the corresponding training raster, but every time I add the explanatory rasters for prediction (they are the same variable but future projections for the same region) the application thinks a bit and then changes the raster back to the original training raster (i.e. it won't allow me to input the prediction raster). I don't know where to go from here because it seems like a bug and nothing I can do on my end.
Please help! Thank you so much in advance for any tips.
Solved! Go to Solution.
If this is the tool, there are lots of inputs, some required and optional. There is no indication that the tool will "revert" a parameter if it doesn't like it.
Could you elaborate on what you used paying particular attention to the paths to the data and the input data types.
Yes that is the tool I am talking about, I've used it previously and had no problems so it is odd that this is happening now. I am using climate data (max temp, precip, wind speed, etc.) as explanatory variables and am having the issue when using the trained model to predict. The data it is rejecting are just future rasters of the projected climate variables that correspond with the variables that the model was trained with (for example, max temp for 2030 under RCP 4.5 scenario). All the data is within the same geodatabase that I have been using for this project and that I have had no issues with prior. The screen capture attached to this reply shows where I am inputting the future corresponding raster data, but the tool is rejecting it.
Thanks for the response.