Selection of the structure of the regression model

Abstract

When constructing regression models, its specific structure is not known a priori in most cases. At the same time, the number of possible theoretical variants of the models is very large. There are no unequivocal theoretical justifications for choosing a model. The available recommendations very often contradict each other when trying to apply them practically. When choosing a marginal (specific) structure of the model, it is suggested to rely on the requirements of the applied problem, which is possible even in the absence of information about the structure.
 Explanations are given on the example of problems on the study of the stability of drills depending on their geometric parameters. Decision-making on both the general structure (the type of formula) and the specific one (the list of model members) is considered. Decision-making is based on the purpose of building the model and the requirements of the applied industry for it.