Case Summary
Forward stepwise regression is a greedy search algorithm that identifies explanatory variables to use in modeling the response in a regression. Stepwise regression is capable of identifying a parsimonious model with few explanatory variables that nonetheless describes the data as well as the saturated model that includes all of the explanatory variables. The stepwise search picks those explanatory variables that obtain a smaller p-value than given by the p-to-enter threshold. We recommend setting this tuning parameter to p-to-enter = 0.05/m, where m is the number of potential explanatory variables. This choice reduces the chance for over-fitting, adding variables to a model that superficially improve the fit while actually degrading ...
Get Statistics for Business: Decision Making and Analysis, 3rd Edition 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.
