Model selection using data mining

Abstract

In this chapter, a new simple procedure for selecting econometric models is provided. It is based on a heuristic approach, called genetic algorithms, which are used to explore the universe of models made available by a general unrestricted model. This search process for the correct model is guided only by the Schwarz information criterion, which acts as the loss function of the genetic algorithm in order to rank the models. This procedure shows good performance relative to other alternative methodologies. A specific example in the world of finance, which shows how to select a tracking portfolio of the IBEX35 Spanish stock market index, is provided.