Book review

To select a model from a family of parametric statistical models one can use different criteria among which the Akaike information criterion (AIC) is one of the most popular. For a given model AIC is defined as twice the difference between the number of free parameters and the maximum of the log likelihood. The AIC approach declares the model minimizing AIC as the most suitable for a given data. The book contains a lot of examples illustrating the usefulness of the AIC approach. The authors belief is that the methods based on AIC can be used in a wide range of applications including the problems of data analysis and model building alike.

The presentation is quite simple and the technique used does not go far from what can be found in an ordinary textbook on mathematical statistics. There are a lot of Fortran programs supplied in an appendix to the book.

The book is written very clearly, it can be used as both a textbook and a reference book and might be very helpful for those applying classical statistical methods in practice.

The present volume was the first systematic book on AIC and it is still the unique one and holds its significance. It looks that still there are no similar books in the literature on mathematical statistics. If not to consider the simplicity of the book as its drawback then I can mention only the following two: its age and quality of English which is not perfect. Purchase of the book may still be a good investment for those dealing with applications of classical statistical methods.