Although the help files are very helpful on showing how to implement the various model fitting procedures, I haven't been able to find a guide on which routine is appropriate in which scenario. Maybe this guide does exist, in which case, please just point me in the right direction. If not, I'd love to see a webinar on examples of fitting with the different available routines and comparing their relative strengths/weaknesses in given scenarios. When is the general Fit Model function appropriate, and when would more specific techniques such as the Gaussian Process or Neural network model appropriate? I'm looking for a relatively high-level overview.
For context, I'm working with a deterministic computer simulation experiment trying to come up with a metamodel for our design/tolerance space.