cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Choose Language Hide Translation Bar

Parameter tuning in Predictive Modeling

In the Predictive Modeling platform, it would be nice to have options, similar to the alternative confusion matrix cutoffs, to change various tuning parameters for a number of iterative steps of values. For example, in a partition, vary the minimum split size between 3 and 10 in increments of 1 (similarly for the other paramters). I realize this is not difficult to accomplish with a loop in a script, but it would be nice if it was as easy as with the confusion matrix cutoff example, or the way it works to put a variable in the By box when doing an analysis. It would then be easy to examine how various performance measures change as the tuning parameters vary.

It has been my experience that most of the default parameters in JMP work quite well, but it becomes tedious to test whether or not this is true. This suggestion would automate this testing and make it easier to determine whether or not to override the defaults.

2 Comments
Nazarkovsky
Level IV

I totally support this proposal. So called hyperparametric tuning frequently encountered in modeling. Well, including this option will make JMP Pro much more competitive to other software for the data analysis, ML modeling like Matlab, for example.

 

 

Status changed to: Acknowledged

Hi @dale_lehman, thank you for your suggestion! We have captured your request and will take it under consideration.