zbMATH — the first resource for mathematics

Akaike information criterion statistics. (Transl. from the Japanese). (English) Zbl 0608.62006
This book presents a practical and unified new approach to statistics based on the Akaike information criterion introduced by H. Akaike in 1971.
An analysis is given of many statistical problems concerning estimation, statistical testing and descriptive methods of data. These are reviewed in a consistent framework of first developing a model then evaluating this by an information measure. This approach obviates the theory of sampling distributions and the need for accompanying statistical tables.
This is the first systematic book in English about the criterion and its uses in data analysis and modeling, and contains all that is needed to use the criterion in practice. In additon, it is an excellent and interesting book.
Reviewer: L.Pardo

62B10 Statistical aspects of information-theoretic topics
62-02 Research exposition (monographs, survey articles) pertaining to statistics
94A15 Information theory (general)
62J05 Linear regression; mixed models
62J10 Analysis of variance and covariance (ANOVA)