Data defines the model by dint of genetic programming, producing the best decile table.


What Criteria Do You Use to Build a Model that Maximizes the Cum Lift?
Bruce Ratner, Ph.D.

The standard criterion for determining which Prediction Model (for which the dependent variable is continuous) among the final set of candidate models is MSE (mean squared error). The standard criterion for determining which Classification Model (for which the dependent variable is categorical, typically binary) among the final set of candidate models is the Classification Table (aka the confusion matrix).

Accurate predictions, and classifications are standard fair for OLS and logistic regression models, respectively. But, in the DM/KD Industry, the models of interest are focused on maximizing the Cum Lift, a measure of model predictiveness/classificativeness of identifying the upper performing individuals often displayed in a decile table. Note: Using OLS and logistic regression models are surrogates for models that seeking optimal Cum Lifts. (Why?) 

What Criteria Do You Use to Build a Model that Maximizes the Cum Lift?

Thanks for sharing.

Bruce_sig



Related Article
What Criteria Do You Use to Determine the Best Model?

For more information about this article, call Bruce Ratner at 516.791.3544 or 1 800 DM STAT-1; or e-mail at br@dmstat1.com.
Sign-up for a free GenIQ webcast: Click here.