3.7. Statistics Measuring Predictive Power
Another class of statistics describes how well you can predict the dependent variable based on the values of the independent variables. This is a very different criterion from the goodness-of-fit measures that we’ve just been considering. It’s entirely possible to have a model that predicts the dependent variable very well, yet has a terrible fit as evaluated by the deviance or the HL statistic. Nor is it uncommon to have a model that fits well, as judged by either of those goodness-of-fit statistics, yet has very low predictive power.
For least squares linear regression, predictive power is usually measured by the coefficient of determination, commonly known as R2. Many different R2 measures have been ...
Get Logistic Regression Using SAS®: Theory and Application now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
