Using JMP® Pro to Build Models Using Generalized Regression Variable Selection Techniques

Presenter: Bill Worley

Fitting, Saving, Comparing and Selecting the Best Predictive Model for Your Circumstances

See how to:

  • Build a response surface model for as many terms as needed
  • Examine Parameter Estimates VIF (Variable Inflation Factor) to identify highly correlated factors
  • Examine Prediction Profiler to identify collinearity
  • Publish prediction formulas to Formula Depot for subsequent model comparison
  • Use validation to avoid model overfitting
  • Determine when and how to use advanced control options, including enforcing effect heredity, forcing terms, and early stopping
  • Use Variable Importance setting to organize Prediction Profiler by the most important factors
  • Build and compare models using Forward Selection (Stepwise), Ridge, Lasso, Double  Lasso, Elastic Net and the adaptive Elastic Net regularization techniques to help identify key explanatory factors

Resources mentioned in these videos

Back to Top