Geisser, Seymour Predictive inference. An introduction. (English) Zbl 0824.62001 Monographs on Statistics and Applied Probability. 55. London: Chapman and Hall, xii, 264 p. (1993). Prediction was the earliest and most prevalent form of statistical inference. However, currently most statistical analyses generally involve inferences or decisions about parameters or indexes of statistical distributions. In view of the author, it is more realistic to channel inferences and decisions in an observational or predictive framework. The primary purpose of this book is an attempt to revive interest in the model of predictive inference about realizable values not observed, based on values that have been observed. The text consists of ten chapters with a bibliography, author index and subject index. The material of the book covers the following topics: non- Bayesian predictive approaches, Bayesian prediction, selecting a statistical model and predicting, problems of comparison and allocation, perturbation analysis, process control and optimization, screening tests for detecting a characteristic, multivariate normal prediction, and interim analysis and sampling curtailment. The book is addressed mainly to statisticians and students in statistics who have at least one year of theoretical statistics, requiring knowledge of liner algebra and advanced calculus. Reviewer: K.Alam (Clemson) Cited in 143 Documents MSC: 62-01 Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics Keywords:prediction; predictive inference; non-Bayesian; Bayesian prediction; comparison; allocation; perturbation analysis; process control; optimization; screening tests; multivariate normal prediction; interim analysis; sampling curtailment PDFBibTeX XMLCite \textit{S. Geisser}, Predictive inference. An introduction. London: Chapman and Hall (1993; Zbl 0824.62001)