Criteria for selection of stochastic linear model selection

Abstract

This research article proposes criteria for the selection of the best stochastic linear regression model. This is one of the three special problems of stochastic linear regression model namely, model selection, misspecification of the model and selection of regressors. Selection of the best model is an important part of stochastic model building. Alarge number of methods have been developed in the literature for selecting best stochastic linear regression model. Y. Tuac et al[7], in 2017, in his research article, presented a small simulation study and real data example to illustrate the performance of the proposed method for dealing with the variable selection and the parameter estimation in restricted linear regression models. Jussi Matta, [6], in his paper, studied model selection methods for two domains linear regression and phylogenetic reconstruction focussing particularly on situations where the amount of data available in either small or very large. Guoqui et al.[9] in their paper presented several model selection criteria which generally can be classified as the penalized robust method. B.M. Potscher, [10], in his research study presented the more general case of regression selection in stochastic linear regression model. Timoterasvirta et al.[8] in his research paper discussed the problem choosing a linear model from a set of nested alternatives.