Henry P. Wynn – författare

This up-to-date account of algebraic statistics and information geometry explores the emerging connections between the two disciplines, demonstrating how they can be used in design of experiments and how they benefit our understanding of statistical models, in particular, exponential models. This book presents a new way of approaching classical statistical problems and raises scientific questions that would never have been considered without the interaction of these two disciplines. Beginning with a brief introduction to each area, using simple illustrative examples, the book then proceeds with a collection of reviews and some new results written by leading researchers in their respective fields. Part III dwells in both classical and quantum information geometry, containing surveys of key results and new material. Finally, Part IV provides examples of the interplay between algebraic statistics and information geometry. Computer code and proofs are also available online, where key examples are developed in further detail.

Written by pioneers in this exciting new field, Algebraic Statistics introduces the application of polynomial algebra to experimental design, discrete probability, and statistics.It begins with an introduction to Gröbner bases and a thorough description of their applications to experimental design. A special chapter covers the binary case with new application to coherent systems in reliability and two level factorial designs. The work paves the way, in the last two chapters, for the application of computer algebra to discrete probability and statistical modelling through the important concept of an algebraic statistical model.As the first book on the subject, Algebraic Statistics presents many opportunities for spin-off research and applications and should become a landmark work welcomed by both the statistical community and its relatives in mathematics and computer science.

Written by pioneers in this exciting new field, Algebraic Statistics introduces the application of polynomial algebra to experimental design, discrete probability, and statistics.It begins with an introduction to Gröbner bases and a thorough description of their applications to experimental design. A special chapter covers the binary case with new application to coherent systems in reliability and two level factorial designs. The work paves the way, in the last two chapters, for the application of computer algebra to discrete probability and statistical modelling through the important concept of an algebraic statistical model.As the first book on the subject, Algebraic Statistics presents many opportunities for spin-off research and applications and should become a landmark work welcomed by both the statistical community and its relatives in mathematics and computer science.

From the Preface: "The theory of optimal design of experiments as we know it today is built on asolid foundation developed by Jack Kiefer, who formulated and resolved some of the major problems of data collection via experimentation. A principal ingredient in his formulation was statistical efficiency of a design. Kiefer's theoretical contributions to optimal designs can be broadly classified into several categories: He rigorously defined, developed, and interrelated statistical notions of optimality. He developed powerful tools for verifying and searching for optimal designs; this includes the "averaging technique"... for approximate or exact theory, and "patchwork"... for exact theory... Kiefer and Wolfowitz provided a theorem now known as the Equivalence Theorem. This result has become a classical theorem in the field. One important feature of this theorem is that it provides a measure of how far a given design is from the optimal design. He characterized and constructed families ofoptimal designs. Some of the celebrated ones are balanced block designs, generalized Youden designs, and weighing designs. He also developed combinatorial structures of these designs. Kiefer's papers are sometimes difficult. In part this is due to the precision and care he exercised, which at times forced a consideration of pathologies and special cases...A reading of his papers on design is replete with examples of his scholarship, his innovativeness, ingenuity, and strength as a researcher."

From the Preface: "Jack Kiefer's sudden and unexpected death in August, 1981, stunned his family, friends, and colleagues. Memorial services in Cincinnati, Ohio, Berkeley, California, and Ithaca, New York, shortly after his death, brought forth tributes from so many who shared in his life. But it was only with the passing of time that those who were close to him or to his work were able to begin assessing Jack's impact as a person and intellect. About one year after his death, an expression of what Jack meant to all of us took place at the 1982 annual meeting of the Institute of Mathematical Statistics and the American Statistical Association. Jack had been intimately involved in the affairs of the IMS as a Fellow since 1957, as a member of the Council, as President in 1970, as Wald lecturer in 1962, and as a frequent author in its journals. It was doubly fitting that the site of this meeting was Cincinnati, the place of his birth and residence of his mother, other family, and friends.Three lectures were presented there at a Memorial Session - by Jerry Sacks dealing with Jack's personal life, by Larry Brown dealing with Jack's contributions in statistics and probability, and by Henry Wynn dealing with Jack's contributions to the design of experiments. These three papers, together with Jack's bibliography, were published in the Annals of Statistics and are included as an introduction to these volumes."

From the Preface: "The theory of optimal design of experiments as we know it today is built on asolid foundation developed by Jack Kiefer, who formulated and resolved some of the major problems of data collection via experimentation. A principal ingredient in his formulation was statistical efficiency of a design. Kiefer's theoretical contributions to optimal designs can be broadly classified into several categories: He rigorously defined, developed, and interrelated statistical notions of optimality. He developed powerful tools for verifying and searching for optimal designs; this includes the "averaging technique"... for approximate or exact theory, and "patchwork"... for exact theory... Kiefer and Wolfowitz provided a theorem now known as the Equivalence Theorem. This result has become a classical theorem in the field. One important feature of this theorem is that it provides a measure of how far a given design is from the optimal design. He characterized and constructed families ofoptimal designs. Some of the celebrated ones are balanced block designs, generalized Youden designs, and weighing designs. He also developed combinatorial structures of these designs. Kiefer's papers are sometimes difficult. In part this is due to the precision and care he exercised, which at times forced a consideration of pathologies and special cases...A reading of his papers on design is replete with examples of his scholarship, his innovativeness, ingenuity, and strength as a researcher."

From the Preface: "Jack Kiefer's sudden and unexpected death in August, 1981, stunned his family, friends, and colleagues. Memorial services in Cincinnati, Ohio, Berkeley, California, and Ithaca, New York, shortly after his death, brought forth tributes from so many who shared in his life. But it was only with the passing of time that those who were close to him or to his work were able to begin assessing Jack's impact as a person and intellect. About one year after his death, an expression of what Jack meant to all of us took place at the 1982 annual meeting of the Institute of Mathematical Statistics and the American Statistical Association. Jack had been intimately involved in the affairs of the IMS as a Fellow since 1957, as a member of the Council, as President in 1970, as Wald lecturer in 1962, and as a frequent author in its journals. It was doubly fitting that the site of this meeting was Cincinnati, the place of his birth and residence of his mother, other family, and friends.Three lectures were presented there at a Memorial Session - by Jerry Sacks dealing with Jack's personal life, by Larry Brown dealing with Jack's contributions in statistics and probability, and by Henry Wynn dealing with Jack's contributions to the design of experiments. These three papers, together with Jack's bibliography, were published in the Annals of Statistics and are included as an introduction to these volumes."

The commentaries in this volume provide reviews of selected papers from the three-volume Collected Papers of Jack Carl Kiefer. From the Preface of Volume III: "The theory of optimal design of experiments as we know it today is built on a solid foundation developed by Jack Kiefer, who formulated and resolved some of the major problems of data collection via experimentation. A principal ingredient in his formulation was statistical efficiency of a design. Kiefer's theoretical contributions to optimal designs can be broadly classified into several categories: He rigorously defined, developed, and interrelated statistical notions of optimality. He developed powerful tools for verifying and searching for optimal designs; this includes the "averaging technique"... for approximate or exact theory, and "patchwork"... for exact theory... Kiefer and Wolfowitz provided a theorem now known as the Equivalence Theorem. This result has become a classical theorem in the field. One importantfeature of this theorem is that it provides a measure of how far a given design is from the optimal design. He characterized and constructed families of optimal designs. Some of the celebrated ones are balanced block designs, generalized Youden designs, and weighing designs. He also developed combinatorial structures of these designs."

Written by pioneers in this exciting new field, Algebraic Statistics introduces the application of polynomial algebra to experimental design, discrete probability, and statistics.It begins with an introduction to Gröbner bases and a thorough description of their applications to experimental design. A special chapter covers the binary case with new application to coherent systems in reliability and two level factorial designs. The work paves the way, in the last two chapters, for the application of computer algebra to discrete probability and statistical modelling through the important concept of an algebraic statistical model.As the first book on the subject, Algebraic Statistics presents many opportunities for spin-off research and applications and should become a landmark work welcomed by both the statistical community and its relatives in mathematics and computer science.

This volume contains the majority of the papers presented at the 5th Inter­ national Workshop on Model-Oriented Data Analysis held in June 1998. This series started in March 1987 with a meeting on the Wartburg, Eisenach (Germany). The next three meetings were in 1990 (St Kyrik monastery, Bulgaria), 1992 (Petrodvorets, StPetersburg, Russia) and 1995 (Spetses, Greece). The main purpose of these workshops was to bring together lead­ ing scientists from ''Eastern'' and ''Western'' Europe for the exchange of ideas in theoretical and applied statistics, with special emphasis on experimen­ tal design. Now that the separation between East and West has become less rigid, this dialogue has, in principle, become much easier. However, providing opportunities for this dialogue is as vital as ever. MODA meetings are known for their friendly atmosphere, leading to fruitful discussions and collaboration, especially between young and senior scien­ tists. Indeed, many long term collaborations were initiated during these events. This intellectually stimulating atmosphere is achieved by limiting the number of participants to around eighty, by the choice of location so that participants can live as a community, and, of course, through the care­ ful selection of scientific direction made by the Programme Committee.

This volume contains a substantial number of the papers presented at the mODa 7 conference. mODastands for Model Oriented Data Analysis and pre- vious conferences have been held in Wartburg (1987) (then in the GDR), St Kirik monastery, Bulgaria (1990), Petrodvorets, St Petersburg, Russia (1992), The island of Spetsos, Greece (1995), the Centre International de Rencontres Mathematiques, Marseilles, France (1998) and Puchberg/Schneeberg, Aus- tria, (2001). The purpose of these workshops has traditionally been to bring together scientists from the East and West interested in the optimal design of exper- iments and related topics and younger and senior researchers in the field. These traditions remain vital to the health of the series. During this period Europe has seen increasing unity and the organizers of and participants in mODa must take some satisfaction from the fact that the youthful ideals of the founders of the series are reflected in this transition. The present conference and mODa 6 are supported by a European Union conference grant (contract HPCF-CT 2000 00045) whose funding emphasis is on younger participants.The company GlaxoSmithKline has very generously continued its support. We are very grateful for these substantial contribu- tions.

The Fifth International Workshop on Model-Oriented Data Analysis (MODA5) focused on experimental design, particularly optimum design. A strength of this series of workshops is that they bring together leading scientists from "Eastern" and "Western" Europe. The proceedings therefore provides a valuable reference to the work of groups from many countries. In addition to 11 papers on optimum designs for linear and nonlinear models, there are groups of papers on designs for quality improvement, designs in agriculture and for model building. Non-design problems include robustness in linear models and estimation problems in nonlinear models. The volume concludes with a discussion on the teaching of experimental design.