under ArcHydro > Wetland Identification Model toolset > there's a tool called `Train Random Trees` which has a parameter called `Output Variable Importance` which returns text file showing how important each band (variable) for the model learning. I think this (Variable Importance) should be adopted under broader deep learning tools as well as it can really help determine which variables are most important while training a Deep Learning model.
Tool Output:
@PavanYadav any thoughts on this?
Connectez-vous pour publier, suivre le contenu, et plus encore. Nouveau ici ? Inscrivez-vous gratuitement.