Practical Predictive Analytics

Goodness of fit statistics

Goodness of fit tests determine how well a model fits the data. There are several methods of judging how well the data fits the logistic regression model:

• Data partitioning: One way we learned how to do this was by dividing the data into training and testing datasets and measuring the degree to which they produce similar results. That method was demonstrated in Chapter 2, The Modeling Process.
• Single metrics: Another way to do this is to come up with a quantitative measure of how well your independent variable predicts your dependent variables on a scale from 0 to 1, with 0 meaning no predictive power, and 1 meaning the best possible predictive power. Area Under the Curve (AUC), and R-square are examples of these ...