Applied data analysis. The Bayesian way (Q400732)

The book under review is aimed at providing a practical guide for the analysis and prognosis of not directly measurable variables based on the Bayesian approach. The book is organized in 14 chapters and one appendix and addresses to undergraduate students in engineering and economic science. Chapter 1 gives an overview of the methods presented in the following chapters (e.g. the rule of Bayes) and shows the application of statistical methods in the framework of quality control. After an introduction to aspects of design of experiments in Chapter 2, the roles of randomisation, replication and controlling are discussed and different types of random sampling are presented in Chapter 3. In Chapter 4, rules for calculating probabilities and computational algorithms for simulation (e.g. Monte Carlo) are provided. Chapter 5 deals with Bayes' rule as well as with Bayesian statistical hypothesis testing. Methods for the assessment of trends and correlations in series of time-dependent observations are presented in Chapter 6. After a short introduction to methods for the prediction of observations in Chapter 7, Chapter 8 illustrates principles of model selection problems (e.g. maximum entropy). In the following two chapters, relevant statistical probability distributions are presented: the exponential and the Poisson distribution in Chapter 9 and the normal distribution in Chapter 10. Chapter 11 provides methods for exploratory data analysis. An introduction to regression models is given in Chapter 12, and Chapter 13 deals with the estimation of parameters of a regression model. Chapter 14 provides a description of the standard error and the comparability of models. The book ends with an appendix with formulas for different data models. In summary, the book under review is recommended as a practical guide for readers interested in Bayesian data analysis. Each chapter of the book is accompanied by numerous practical examples from engineering, economic and medical science as well as by a huge set of reflection questions. Nevertheless, the reader would benefit if solutions to the questions would be provided and the implementation of the presented methods would be illustrated by means of a software package.