What inspired this wish list request?
We have numbers of processes that need mixed multi-variate modeling and understanding.
Currently we have an Excel and Python implementation of the following paper:
https://besjournals.onlinelibrary.wiley.com/doi/10.1111/j.2041-210x.2012.00261.x
To estimate fixed parameter variance, random parameters variance, and the residual and hence calculate R-squared values for each.
JMP 17.2 does not do this automatically hence Excel and Python implementations.
What is the improvement you would like to see?
Please see the attached mark-up (an output from our Python script) that directly marks the R2 on the marginal (fixed) effects plot, marginal + conditional effects plot. It also numerically displays marginal, conditional, delta between conditional and marginal and the residual as a table (not illustrated).
Why is this idea important?
Most real-world process engineers think in terms of R-squared values. We have > 130 engineers doing this manually in excel, which is now pythonized. However a directly solution in JMP will simplify our life significanlty and reduce human errors. It will also save us a lot of time that now gets invested in Python installations etc.
Financial impact of this idea and why we need it ASAP (despite internal Python / Excel implementation):
Estimated at > 1M US$/month. Robust statistical analysis is an important enabler. Having this capability directly in JMP is very important. Hence this request.
Other background on JMP community:
There was also a robust discussion on JMP community discussion on this as below:
https://community.jmp.com/t5/Discussions/Generating-conditional-and-marginal-R-squared-for-mixed-mod...
