An extension of regression trees to generate better predictive models
Kim, H., F. M. Guess, and T. M. Young.  2011.  IIE Transactions, 43(1):43-54. doi: 10.1080/0740817X.2011.590441.

For situations where the data are drawn from reasonably homogeneous populations, traditional methods such as multiple regression typically yield insightful analyses. For situations where the data are drawn from more heterogeneous populations, decision tree approaches, such as CART (Classification And Regression Trees) and GUIDE (Generalized, Unbiased, Interaction Detection and Estimation), are more likely to recognize idiosyncratic subpopulations and interactions automatically. In contrast to CART, however, GUIDE yields models with better predictive performance for each subpopulation. We extend the idea of GUIDE to handle ANCOVA (Analysis of Covariance) type problems. This paper compares GUIDE modeling to various decision tree methods and to multiple regression. The paper identifies and discusses the relative advantages and disadvantages of multiple regression, CART, and GUIDE. GUIDE produces quality or reliability models that exhibit greater predictive accuracy than multiple regression or CART for complex, highly diverse populations. Also, GUIDE is readily applicable to many other areas, such as repairability and maintainability settings involving both qualitative and quantitative variables. We illustrate GUIDE with a small case study of an engineered wood product, medium density fiberboard.