We have a Pro prospect who wants to know if JMP offers partial dependence plots (PDPs) for assessing main effects in machine learning models (particularly tree-based models). I'm familiar with our Marginal Model Plots (MMPs) provided with the Profiler's Assess Variable Importance, and from what I can Google about PDPs, they are equivalent in interpretation to our MMPs although perhaps calculated differently (i.e. PDPs use some sort of grid of predictor values whereas our MMPs use Monte Carlo methods). Does anybody know if I'm on (or off) here and if there are any other differences to be aware of? Thanks in advance.
