Statistical analysis of panel count data (Q368416)

The book under review presents on a total of 271 pages an introduction into the methodology of analysing panel count data. It addresses to scientists and graduate students with basic knowledge about probability theory and statistics. The main focus of the monograph is on the methodology, nevertheless some applications to real data are provided. Each chapter is accompanied by a detailed paragraph providin a bibliography, remarks and discussion of the presented methods. In the Introduction (Chapter 1) the authors present examples of different panel count data, basic notations and concepts about counting processes and analysis of recurrent event data and panel count data. Before considering semi- and nonparametric models, the most commonly used parametric approach of Poisson models are introduced in Chapter 2. The following Chapters 3 and 4 are devoted to the nonparametic estimation (Chapter 3) and nonparametic two-sample comparison of point processes (Chapter 4). Regression analysis of panel count data assuming independence of the observation process and the underlying recurrent event process given covariate processes is described in Chapter 5, whereas in Chapter 6 regression analysis of panel count data is presented for situations where this assumption may not hold and the observation process depends on the recurrent event process. Chapter 7 is devoted to the extension to multivariate panel count data with several related recurrent event processes and basic methodology on nonparametric treatment comparisons and semiparametric regression analysis is presented. Chapter 8 provides additional methods on analysis of panel count data like variable selection models, analysis of mixed recurrent event and panel count data, the analysis of panel count data arising from multi-state models and Bayesian approaches for analysis of panel count data. The volume ends with additional information on the data sets used throughout the monograph (Chapter 9). In summary, the book under review is recommended to researchers with strong background of probability and statistics interested in methodology on panel count data.