Information criteria and statistical modeling. (Q2460366)

With the main purpose of explaining the critical role of information criteria in statistical modeling, this book is written by two leading experts. Explaining the basic concepts of statistical modeling in Chapter 1, the authors describe probability distribution and conditional distribution models in Chapter 2. Chapter 3 discusses Kullback-Leibler information as a criterion for statistical model selection and how this leads to the Akaike Information Criterion (AIC). Chapter 4 discusses the use of AIC in statistical inference problems like testing for equality of distributions, determining the bin size of a histogram, selecting the order of a regression model, and so on. Generalized Information Criterion (GIC) is the title of Chapter 5. GIC was introduced by the authors as a general framework for constructing information criteria for functional statistics which employ estimation procedures other than the method of maximum likelihood. GIC is introduced and relationships among various information criteria are discussed. Applications of GIC in nonlinear modeling and model evaluation are discussed in Chapter 6. Chapter 7 introduces a general framework for constructing information criteria in the context of functional statistics and gives a detailed derivation of the GIC. Asymptotic properties of information criteria are also discussed in this chapter. Chapter 8 discusses the bootstrap information criterion and its applications. Chapter 9 is on Akaike's Bayesian Information Criterion. Chapter 10, the last chapter, discusses several model evaluation criteria such as cross-validation, generalized cross-validation, final predictive error, Mallow's \(C_p,\) the Hannan-Quinn criterion, and informational complexity (ICOMP). The book ends with a list of references and an index. The style of writing is very good. Examples illustrate the concepts discussed and make the book immensely readable. The book achieves its objective of providing comprehensive explanations of the concepts and derivations of several information criteria like the AIC, BIC, GIC, and so on. Anybody interested in statistical modeling will love to read this book. There are very few misprints and it will be very useful to researchers and students interested in learning statistical modeling and model evaluation.