Practical Predictive Analytics

Variable selection

Stepwise regression can be used as a variable selection method in cases in which you have a large number of variables, and you need to define a starting point in which you can start looking at a limited set of variables. However, all of the caveats regarding p-values come into play here as well. Use this method only to look at a candidate list of variables for model testing. It is not a good idea to use stepwise regression to determine a final model, since there are several methods use to determine the best variables, the final model can yield artificially high and biased results.

However, let's just start with identifying a list of potential variables for our PainGLM model, which is small, but will be good to illustrate ...