This website uses Cookies. Click Accept to agree to our website's cookie use as described in our Privacy Policy. Click Preferences to customize your cookie settings.
Accept
Reject
Preferences
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: 
Choose Language Hide Translation Bar
tonya_mauldin
Staff tonya_mauldin
Staff
Modernizing the Distribution platform – New fitters in JMP 15
Created: Nov 13, 2019 08:55 AM

The Distribution platform is one of the most widely used platforms in JMP. This platform is not only used for testing which distribution fits your data, it is used for data exploration, capability analysis, and so much more. Distribution has been around since the first version of JMP, and it was time for some changes. With JMP 15 comes the modernization of this commonly used platform. 

New Fitters

One of the biggest changes in the Distribution platform is related to fitting distributions. The fitters have been modernized to use the same code as the Generalized Regression platform. Sharing code between JMP platforms in this way ensures consistency and requires less maintenance going forward. New Distribution fits were added while we were updating this code.

You can now fit a Negative Binomial distribution (equivalent to the Gamma-Poisson distribution), a Cauchy distribution, a ZI (zero-inflated) Poisson distribution, and a ZI (zero-inflated) Negative Binomial distribution. The Johnson fits are now a single command that fits the best-fitting distribution from the Johnson system of distributions. This Johnson fitting method is the same method (quantile matching) used in the Process Capability platform. This method is more stable and faster than maximum likelihood.

ContsFits.pngDiscreteFits.png

To see an example of the new Johnson fitter, open Airline Delays.jmp found in the sample data folder.  

Open("$SAMPLE_DATA/Airline Delays.jmp");

Select Analyze->Distribution.  Specify Arrival Delay as Y, Columns and click OK.

dialog.png

Click on the red triangle next to Arrival Delay and select Continuous Fit->Fit Johnson. The best fit from the family of Johnson distributions for this data is the Johnson Su Distribution. 

Johnsonex.png

Prior to JMP 15, the only way to control the Kernel Std for the Nonparametric Density was via a slider using the interface. Now, there is the option to specify a specific value for the bandwidth parameter. JMP 15 also gives you the ability to script the bandwidth parameter.

nonparametric.png

Before JMP 15, no standard errors were given for the parameter estimates for the SHASH distribution. This has been added in JMP 15 to be consistent with the other distributional fits.

SHASH.png

Summary

  • New fitters that match the Generalized Regression platform.
  • Cauchy Distribution.
  • Negative Binomial Distribution.
  • ZI Poisson Distribution.
  • ZI Negative Binomial Distribution.
  • Single Johnson fitter that shows the best Johnson fit.
  • Ability to enter a value for the smooth curve bandwidth parameter (also the ability to script this parameter).
  • Parameter estimate standard errors for the SHASH distribution.

This blog post only scratches the surface of the JMP 15 new features in the Distribution platform. Look for my next blog post in which I detail the new features for comparing distributional fits.

Last Modified: Nov 13, 2019 11:08 AM