Model Selection in Beta Regression Analysis Using Several Information Criteria and Heuristic Optimization

Authors : Emre DÜNDER, Mehmet Ali CENGİZ

Pages : 76-84

View : 17 | Download : 5

Publication Date : 2020-12-31

Article Type : Research Paper

Abstract :In the context of generalized linear modelling insert ignore into journalissuearticles values(GLM);, the beta regression analysis is used to estimate regression models when the dependent variable lies between insert ignore into journalissuearticles values(0,1);. In this paper, we carried out a model selection process using several information criteria with heuristic optimization. We employed the differential evolution algorithm as a heuristic optimization method to select the best model for beta regression analysis. The results show that the alternative-type information criteria provide competitive results during the model selection process in beta regression analysis.
Keywords : Beta regression, differential evolution algorithm, information criteria, model selection