I need some information and opinions on rescaling and/or centering my predictor varibles.
Part 1 of my question: I have reviewed two basic statistics texts and didn't find enough information. Can anyone recommend a very thorough discussion on standardizing, normalizing, rescaling, centering, etc. in data mining type problems? (I use a variety of data mining algorithms, including regression.) Part 2: I expect my regression models to have high order interaction terms that are significant when the main effects are not signficant. If this were *not* the case, I would center my predictor variables at zero. With this "unorthodox" model form, my intuition tells me I will get better results if I use a range that does not include zero. With high order interactions and zero-centered predictors, if one variable value is equal to the mean, the whole interaction term would be zero. This doesn't model the problem correctly. Any thoughts? FYI, the regression models with high order interaction terms that are significant when the main effects are not signficant are expected to be an intermediate model, not a final deployed model. The final model is expected to be more traditional. David . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
