統計的モデリングと情報量規準

抄録

The problem of evaluating the goodness of statistical models is fundamental and of importance in various fields of statistics, natural sciences, neural networks, engineering, economics, etc. Akaike's . information criterion, known as AIC, provides a useful tool for constructing statistical models, and a number of successful applications of AIC in statistical data analysis have been reported. AIC is a criterion for evaluating the models estimated by the maximum likelihood method. With the development of various non-linear modeling techniques, the construction of criteria which enable us to evaluate various types of statistical models has been required. The aim of this paper is to give a systematic account of some recent developments in model evaluation criteria from information-theoretic and Bayesian points of views. We intend to provide a basic expository account of the fundamental principles behind information criteria. We also discuss the application of the bootstrap methods in model evaluation problems.