Fitting Linear Models • Generalized Regression Models
Previous
 • 
Next
Generalized Regression Models
Build Models Using Variable Selection Techniques
The Generalized Regression personality of the Fit Model platform is available only in JMP Pro.
In JMP Pro, the Fit Model platform’s Generalized Regression personality provides variable selection techniques, including shrinkage techniques, that specifically address modeling correlated and high-dimensional data. Two of these techniques, the Lasso and the Elastic Net, perform variable selection as part of the modeling procedure.
Large data sets that contain many variables typically exhibit multicollinearity issues. Modern data sets can include more variables than observations, requiring variable selection if traditional modeling techniques are to be used. The presence of multicollinearity and a profusion of predictors exposes the shortcomings of classical techniques.
Even for small data sets with little or no correlation, including designed experiments, the Lasso and Elastic Net are useful. They can be used to build predictive models or to select variables for model reduction or for future study.
The Generalized Regression personality is useful for many modeling situations. This personality enables you to specify a variety of distributions for your response variable. Use it when your response is continuous, binomial, a count, or zero-inflated. Use it when you are interested in variable selection or when you suspect collinearity in your predictors. More generally, use it to fit models that you compare to models obtained using other techniques.
Figure 5.1 The Solution Path for an Elastic Net Fit
Contents 
Overview of the Generalized Regression Personality
Example of Generalized Regression
Launch the Generalized Regression Personality
Distribution
Generalized Regression Report Window
Generalized Regression Report Options
Model Launch Control Panel
Estimation Method Options
Advanced Controls
Validation Method Options
Early Stopping
Go
Model Fit Reports
Regression Plot
Model Summary
Estimation Details
Solution Path
Parameter Estimates for Centered and Scaled Predictors
Parameter Estimates for Original Predictors
Active Parameter Estimates
Effect Tests
Model Fit Options
Statistical Details for the Generalized Regression Personality
Statistical Details for Estimation Methods
Statistical Details for Advanced Controls
Statistical Details for Distributions

Previous
 • 
Next
Help created on 10/11/2018