Read more
This book presents the basic ideas of statistical methods in the design of optimal experiments. This new edition now includes sections on design techniques based on the elemental Fisher information matrices (as opposed to Pearson information/moment matrices), allowing a seamless extension of the design techniques to inferential problems where the shape of distributions is essential for optimal design construction. Topics include designs for nonlinear models, models with random parameters and models with correlated observations, designs for model discrimination and misspecified (contaminated) models, and designs in functional spaces.
The authors avoid technical details, assuming a moderate background in calculus, matrix algebra, and statistics. In many places, however, suggestions are made as to how the ideas presented in this book can be extended and elaborated for use in real scientific research and practical engineering problems.
List of contents
Preface.- Introduction.- 1. Some Facts From Regression Analysis.- 2. Convex Design Theory.- 3. Numerical Techniques.- 4. Optimal Design under Constraints.- 5. Special Cases and Applications.- A. Elemental Fisher Information.- B. Selected Formulas from Matrix Theory.- C. List of Symbols.- References.- Index.
About the author
Valerii V. Fedorov is an independent consultant. During the last two decades, he has been involved in the development of optimal design methods for pharmaceutical research; he was Vice President, Innovation Center, ICON plc; Vice President and Head of Predictive Analytics, Quintiles; Head of the Research Statistics Unit, GlaxoSmithKline inc. Professor Fedorov is the author of over 200 publications, including several books. His monograph The Theory of Optimal Experiments (1972, Academic Press) is one of the first books on optimal experimental design; his Optimal Design for Nonlinear Response Models (co-authored with Dr. Sergei Leonov, 2014, CRC) describes methods developed for pharmaceutical research. Valerii is an ASA Fellow, an Honorary Professor of Cardiff University, UK, an Adjunct Scholar of the University of Pennsylvania, USA, and an elected member and former Council member of the International Statistical Institute.
Peter Hackl, born in Linz, Austria, is Professor of Statistics at the Vienna University of Economics and Business. He is an internationally recognized expert in statistical methods and applications and has published on topics such as time series analysis, econometric methods, optimal design of experiments, as well as for statistical methods for process control, quality improvement, and customer satisfaction measurement. He is the author of five books, editor of two, and has published more than 120 articles in refereed scientific journals.
Summary
This book presents the basic ideas of statistical methods in the design of optimal experiments. This new edition now includes sections on design techniques based on the elemental Fisher information matrices (as opposed to Pearson information/moment matrices), allowing a seamless extension of the design techniques to inferential problems where the shape of distributions is essential for optimal design construction. Topics include designs for nonlinear models, models with random parameters and models with correlated observations, designs for model discrimination and misspecified (contaminated) models, and designs in functional spaces.
The authors avoid technical details, assuming a moderate background in calculus, matrix algebra, and statistics. In many places, however, suggestions are made as to how the ideas presented in this book can be extended and elaborated for use in real scientific research and practical engineering problems.
