S-N Curve Fitting - is there a recommended JMP solution?

Question

I have a customer who is characterizing metal alloys with SN testing and he is interested in generating curve fits, determining theoretical material constants and confidence intervals for these constants.  Non-linear platform works beautifully except some of the data is censored.  Has anyone encountered this issue before and is there a recommended solution?

DonMcCormack · Answer

You need to know the the liklihood function. You can get this from the pdf or cdf of the distribution being modeled. There's an example in the Nonlinear chapter in the docs.

Mark_Bailey · Answer

The Nonlinear platform is really a MLE platform. You should be able to write a custom Loss function to accommodate the censoring versus exact data cases with an indicator function. I don't know how, though, 'off the top of my head.'