Statistical modeling in GLIM 4.
2nd ed.

*(English)*Zbl 1078.62001
Oxford Statistical Science Series 32. Oxford: Oxford University Press (ISBN 0-19-852413-7/hbk). xiv, 557 p. (2005).

[For the review of the first edition from 1989 see Zbl 0676.62001.]

The aim of this book is twofold: to provide an exposition of the principles of statistical modeling with the necessary statistical theory, and to describe the application of these principles to the analysis of a wide range of practical examples using GLIM 4, the extended and substantially improved version of GLIM (Generalized Linear Interactive Modeling) statistical package. This second edition of the book is extended in scope and restructured contents, enclosing three new chapters (compared to the first edition) on finite mixtures and random effects models, with detailed coverage of Gaussian quadrature and nonparametric maximum likelihood analysis of these models.

Designed mainly for intensive courses, the book can be used in various manners for graduate and advanced undergraduate courses in statistics, in conjunction with interactive use of GLIM either in lectures or in practical / tutorial sessions. This book is equally suitable as a self-teaching manual for professional statisticians and research workers.

GLIM (with its most recent release, GLIM 4) is a command driven package devoted to statistical modeling as an interactive process. The interactivity between the data and the data modeler is an essential feature of the statistical modeling process. This book focuses on generalized linear models and extensions of these models, which can be fitted using an interactive reweighed least squares (IRLS) algorithm. The computational extensions of GLIM 4 cover various areas, such as survival analysis and random effects models, developed within the framework of the IRLS algorithm that is incorporated into more complex procedures and macros. The command set of GLIM 4 is compact and can be learnt relatively quickly.

GLIM 4 commands for data manipulation, data transformation, data display and model fitting may be entered, with a few restrictions, in any order. In addition, sequences of frequently used commands (known also as GLIM macros) may be named, saved, and used later as the user’s own procedures. While GLIM is primarily designed for interactive use, it can also be run in the batch mode, by submitting a file of previously assembled commands. This is an ideal book for graduate and research students in applied statistics to a wide range of fields, including biology, medicine and social sciences.

For a closer look into the book contents, here are the titles (only) of the enclosed chapters of the book: Chap. 1: Introducing GLIM 4. Chap. 2: Statistical Modeling and Inference. Chap. 3: Regression and Analysis of Variance. Chap. 4: Binary Response Data. Chap. 5: Multinomial and Poisson Response Data. Chap. 6: Survival Data. Chap. 7: Finite Mixture Models. Chap. 8: Variance Components Models.

The aim of this book is twofold: to provide an exposition of the principles of statistical modeling with the necessary statistical theory, and to describe the application of these principles to the analysis of a wide range of practical examples using GLIM 4, the extended and substantially improved version of GLIM (Generalized Linear Interactive Modeling) statistical package. This second edition of the book is extended in scope and restructured contents, enclosing three new chapters (compared to the first edition) on finite mixtures and random effects models, with detailed coverage of Gaussian quadrature and nonparametric maximum likelihood analysis of these models.

Designed mainly for intensive courses, the book can be used in various manners for graduate and advanced undergraduate courses in statistics, in conjunction with interactive use of GLIM either in lectures or in practical / tutorial sessions. This book is equally suitable as a self-teaching manual for professional statisticians and research workers.

GLIM (with its most recent release, GLIM 4) is a command driven package devoted to statistical modeling as an interactive process. The interactivity between the data and the data modeler is an essential feature of the statistical modeling process. This book focuses on generalized linear models and extensions of these models, which can be fitted using an interactive reweighed least squares (IRLS) algorithm. The computational extensions of GLIM 4 cover various areas, such as survival analysis and random effects models, developed within the framework of the IRLS algorithm that is incorporated into more complex procedures and macros. The command set of GLIM 4 is compact and can be learnt relatively quickly.

GLIM 4 commands for data manipulation, data transformation, data display and model fitting may be entered, with a few restrictions, in any order. In addition, sequences of frequently used commands (known also as GLIM macros) may be named, saved, and used later as the user’s own procedures. While GLIM is primarily designed for interactive use, it can also be run in the batch mode, by submitting a file of previously assembled commands. This is an ideal book for graduate and research students in applied statistics to a wide range of fields, including biology, medicine and social sciences.

For a closer look into the book contents, here are the titles (only) of the enclosed chapters of the book: Chap. 1: Introducing GLIM 4. Chap. 2: Statistical Modeling and Inference. Chap. 3: Regression and Analysis of Variance. Chap. 4: Binary Response Data. Chap. 5: Multinomial and Poisson Response Data. Chap. 6: Survival Data. Chap. 7: Finite Mixture Models. Chap. 8: Variance Components Models.

Reviewer: Neculai Curteanu (Iaşi)

##### MSC:

62-04 | Software, source code, etc. for problems pertaining to statistics |

62-01 | Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics |

62J12 | Generalized linear models (logistic models) |

62J10 | Analysis of variance and covariance (ANOVA) |

62N05 | Reliability and life testing |