Using Prediction Profiling to Maximize Model Proficiency – Part 2

See how to:

• ID new JMP 17 capabilites, including JMP extrapolation control
• Customize Profiler appearance
• Share Profilers as HTML
• Use the Interaction Profiler to show or hide interaction plots that update as you update the factor values in the Prediction Profiler
• Use the Surface Profiler to produce a surface plot for the fitted model
• Use the Simulator in the profilers to define random inputs, run simulations, and produce output tables of simulated values.
• Use the Design Space Profiler to to explore the impact of factor limits on the expected rate of future runs passing response specifications (in-spec rate) 
• Examine Contour Profiler plots of response contours for two factors at a time

Questions answered by @andreacoombs1  at the live webinar:

Q: When we were resetting the factor grid, did the numbers 41 indicate 41 levels for each factor?

A; Yes.

Q: How do you determine SD for the simulation?

A:  When you're doing simulation, standard deviation comes up in  different ways.  You have to determine  what standard deviation to use when interjecting  random noise around each of your factors when doing the simulation. JMP proposes standard deviations, but you, as a subject matter expert, can modify them. For example, you may know about some ability in terms of temperature control on your equipment and want to to access process knowledge and some historical data to help you determine what SD to choose.

Q: How do you change the properties before sharing the profiler?

A:  Alt/Click on the red triangle pops up all properties.  Alt/Click works on all JMP reports.  Quick Reference Guide contains many shortcuts like this for Mac and Windows.

Q: Can Gaussian Process model in JMP handle categorical factors?

A: No, neither JMP nor JMP Pro or Standard will categorical Y response.   @PatrickGiuliano  clarified this after the session.  However, X assigns the columns to use as explanatory variables for Gaussian Process.  Categorical variables are allowed in JMP Pro when the Fast GASP option is specified. 

Resources

• Using Prediction Profiling - Part 1
• Mastering video and Q&A on Design Space Profiling
• Prediction Profiler documentation
• Prediction Profiler options, including JMP Pro Extrapolation Control
• Statistical Details on Extrapolation Control in Prediction Profiler