Model choice is a subject that involves artistic and individual components that depend on the area of application, the amount of knowledge on the available models, and one’s sense of aesthetics.
The first author’s interest in model choice began in the summer of 1973: after a year of attending lectures at Imperial College, he made an appointment with his supervisor to decide his thesis topic. They summarized the meeting in the notes shown below. Although he did not pursue exactly the applications discussed at that time, he developed other results and applications concerning separate or nonnested model choice. The beginning
vii Being a Bayesian, the second author began to be interested in the choice of models when solving a problem related to pollution in an industrial city in Brazil.
He applied Bayesian significance tests to the mixture of models proposed by Cox instead of hypothesis tests and discrimination using Bayes factors. Both authors have been following the advances in the subject, and this book is the result of their attempts to do so. The authors are grateful to the writers and researchers on the subject from whom they have benefited and whom they have followed while writing this work, especially Mohammed Hashem Pesaran, and to Annibal P. SantAnna and Marlos Augusto G. Viana, who offered many suggestions for and corrections of the manuscript. The authors are also thankful for the many important contributions of Maria Ivanilde S. Araujo, Edilson F. de Arruda, Cachimo C. Assane, Rodrigo A. Collazo, Marcelo Lauretto, Brian A.R. de Melo, Fernando Poliano, and Julio Stern. They also thank Marcelo Fragoso and Augusto C.G. Vieira for the opportunity to complete their writing at Laboratório Nacional de Computação Científica–LNCC in Petrópolis, Brazil. Evelyn Best and Veronika Rosteck of Springer have been supportive and patient editors.
Petrópolis, Brazil Basilio de Bragança Pereira
February 2016 Carlos Alberto de Bragança Pereira
Click HERE and check out the preface and contents.