Breakthroughs in Statistics: Volume III (Springer Series in Statistics / Perspectives in Statistics)

图书介绍

This is the third volume of a collection of seminal papers in the statistical science written during the past 110 years. These papers have each had an outstanding influence on the development of statistical theory and practice over the last century. Each paper is preceded by an introduction written by an authority in the field providing background information and assessing its influence. Volume III concertrates on articles from the 1980's while including some earlier articles not included in Volumes I and II. Samuel Kotz is Professor of Statistics in the College of Business and Management at the University of Maryland. Norman L. Johnson is Professor Emeritus of Statistics at the University of North Carolina. Also available: Breakthroughs in Statistics Volume I: Foundations and Basic Theory Samuel Kotz and Norman L. Johnson, Editors 1993. 631 pp. Softcover. ISBN 0-387-94037-5 Breakthroughs in Statistics Volume II: Methodology and Distribution Samuel Kotz and Norman L. Johnson, Editors 1993. 600 pp. Softcover. ISBN 0-387-94039-1

Review:
Volume I of breakthroughs in statistics covered the foundations and the schools of inference that developed. Volume II did a good job of covering further developments in methodology and distribution theory. However after the two volume were published in 1992 it became apparent that there had been some key omissions and that further developments shed light on the importance of articles that had not been considered breakthrough papers earlier but in hindsight they clearly were. So in 1997 these fine editors published Volume III to fill in the gaps.

The earliest paper that was overlooked was Galton's famous 1889 paper that introduced regression, correlation and the phenomena of regression toward the mean. It is clear from reading the book that all the added papers were worthy of the title of breakthrough. Of particular interest to me are the articles that deal with faster computing techniques and computer intensive algorithms that were made practically important as the speed of computers grew astronomically in the 1990s. So the paper by Metropolis et al. 1953 became a breakthrough when another breakthrough paper by Geman and Geman 1984 showed the these Markov Chain Monte Carlo algorithms were useful in a Bayesian approach to image processing or "image restoration." After this breakthrough modifications of the basic algorithms have been developed to solve many of the previously viewed as infeasible problems in Bayesian inference. The practical development of frequency domain times series owes a debt to the fundamental breakthrough paper by Cooley and Tukey in 1965 that introduced the first versions of the fast Fourier transform. I shall skip the others except for two of my favorites, Peter Hall's 1988 paper that established theoretically through the use of Edgeworth and Cornish-Fisher expansions the asymptotic accuracy of various types of bootstrap confidence intervals and the 1986 paper by Liang and Zeger that developed an approach to longitudinal data analysis via generalized linear models. In this remarkable paper they also introduced Generalized Estimating Equations that are commonly used today. this work led to a famous and popular book on longitudinal data analysis that they coauthored with Peter Diggle. The pharmaceutical industry that I work in owes a great debt to Liang and Zeger as their approach to longitudinal (or repeated measures) data has been applied to the efficacy endpoints of phase III clinical trials.