Prediction profiler with nominal variables

[#1]

Hi all!

I have 6 continuous variables (N0 to N5) and one nominal variable (named "recode") that I would like to analyze with the prediction profiler. The nominal variable "recode" has 4 values (1,2,3,4).

When I initiate the prediction profiler, the nominal variable is "broken" into three sections (see image below).

How do I interpret it? for example, how to I set the profiler to show the predicted values for all the variables when "recode" equals 1 ?

Thank you!!

[#2]

Hi @Lavik17 : Not sure without knowing more about the data and your proposed model; but, based on what I see I'm guessing you have some singularity issue that can occur when you try to estimate more parameters than you have degrees of freedom. My guess is that it stems from interaction terms that are not estimable (because factor combinations required to estimate interactions are not available). Singularity issues will be identified in the output. Can you share more info about your data and model? 

[#3]

Sorry, I'm a bit confused by your query.  Is it possible for you to attach your experiment (ok to code it).  It appears you may have the categorical (nominal) factor setup incorrectly.  What do you mean by "analyze with prediction profiler"?  You might first want to determine if the factors in your model are useful and statistically significant before looking at the profiler (for example, it does not appear as N2 or N4 have much effect).

[#4]

Is this model from the Stepwise platform?

[#5]

Hi @Lavik17,

As implicitly suggested by @Mark_Bailey depending on the platform you use and the parameters you have used when launching a platform, categorical effects are handled differently :

• For "Fit Model" (Standard Least Squares) approach : 

"When you enter a column with a nominal modeling type in the Fit Model launch window, JMP represents it internally as a set of continuous indicator variables. Each variable assumes only the values –1, 0, and 1. (Note that this coding is one of many ways to use indicator variables to code nominal variables.) If your nominal column has n levels, then n–1 of these indicator variables are needed to represent it. (The need for n–1 indicator variables relates directly to the fact that the main effect associated with the nominal column has n–1 degrees of freedom.) Full details are covered in Nominal Factors."

From : Statistical Details for Nominal Effects Coding

• For "Generalized Regression" (Standard Least Squares or other estimation methods) :

"The parameterization of nominal variables used in the Generalized Regression personality differs from their parameterization using other Fit Model personalities. The Generalized Regression personality uses indicator function parameterization. In this parameterization, the estimate that corresponds to the indicator for a level of a nominal variable is an estimate of the difference between the mean response at that level and the mean response at the last level. The last level is the level with the highest value order coding; it is the level whose indicator function is not included in the model."

From :  Launch the Generalized Regression Personality

So you may end up with different estimates and p-values of effects depending on the platform used.

And finally, if you're using Stepwise method, you may end up with the profiler you have if you don't select the rule "Whole effect", as the Stepwise selection will only select significant terms based on the rules you have specified, which may select only few "levels" from a categorical factor, resulting in the splitting of this categorical factor in several factors on the profiler.

More infos here : https://www.jmp.com/support/help/en/17.2/index.shtml#page/jmp/stepwise-regression-control-panel.shtm...

Hope this will help you,