User Community
cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Show  only  | Search instead for 
Did you mean: 
  • Instantly extract effect sizes, F-ratios, and FDR-adjusted p-values from your models with the Calculate Effects Sizes extension, available now in the JMP Marketplace!
  • New to JMP? Join us Sept. 23-24 for the Early User Edition of Discovery Summit, tailor-made for new users. Register now for free!
  • See how to use the JMP Marketplace – Free tools to expand JMP capabilities. Register. July 10, 2 pm US Eastern Time.

JMP Wish List

We want to hear your ideas for improving JMP. Share them here.
Choose Language Hide Translation Bar

cross validation in Discriminant analysis & other options

Status: Archived Submitted by on ‎12-15-2017 02:26 AM
2 Comments (2 New)

I tried the demo and this is what I really missed in the software

 

- k - fold cross validation and LOO cross validations in Discriminant analysis.

- Display RMSE value  and other indicators (AICc BIC) to assess the model would be nice.

- Give the possibility to randomly select training versus test dataset  (not only in JMP Pro but also in standard soft).

- Also, I noticed we can't change the color of the ellipse (confidence ellispe) of each group, which is an anoying issue when you use custom colors for each group as they don't fit with the ellipse color.  (but I loved the options for markers' shape/color).

- Last, please make the graph menu bar (to save / export your canonical plot and other graph) appears by default. Took me hours to find how to do it and export a graph.

 

Thanks!

2 Comments
Ryan_Gilmore
Community Manager Community Manager
Community Manager
‎10-26-2022 04:37 PM
‎10-26-2022 04:37 PM
Status changed to: Archived
We are archiving this request. If this is still important please comment with additional details and we will reopen. Thank you!
hengehold_da1
Level III
‎11-25-2024 01:32 PM
‎11-25-2024 01:32 PM

I'd like to second the request for cross validation or leave one out validation as part of some kind of stepwise variable selection.  In addition, it would be very helpful if you would create the "Fisher's classification function coefficients" that could be easily used to determine the predicted group (e.g. as part of an online survey to screen/qualify panelists).