Recursive Partitioning and Applications

Description

The routes to many important outcomes including diseases and ultimately death as well as financial credit consist of multiple complex pathways containing interrelated events and conditions. We have historically lacked effective methodologies for identifying these pathways and their non-linear and interacting features. This book focuses on recursive partitioning strategies as a response to the challenge of pathway characterization. A highlight of the second edition is the many worked examples, most of them from epidemiology, bioinformatics, molecular genetics, physiology, social demography, banking, and marketing. The statistical issues, conceptual and computational, are not only treated in detail in the context of important scientific questions, but also an array of substantively-driven judgments are explicitly integrated in the presentation of examples. Going considerably beyond the standard treatments of recursive partitioning that focus on pathway representations via single trees, this second edition has entirely new material devoted to forests from predictive and interpretive perspectives. For contexts where identification of factors contributing to outcomes is a central issue, both random and deterministic forest generation methods are introduced via examples in genetics and epidemiology. The trees in deterministic forests are reproducible and more easily interpretable than the components of random forests. Also new in the second edition is an extensive treatment of survival forests and post-market evaluation of treatment effectiveness. Heping Zhang is Professor of Public Health, Statistics, and Child Study, and director of the Collaborative Center for Statistics in Science, at Yale University. He is a Fellow of the American Statistical Association and the Institute of Mathematical Statistics, a Myrto Lefkopoulou Distinguished Lecturer Awarded by Harvard School of Public Health, and a Medallion lecturer selected by the Institute of Mathematical Statistics. Burton Singer is Courtesy Professor in the Emerging Pathogens Institute at University of Florida, and previously Charles and Marie Robertson Professor of Public and International Affairs at Princeton University. He is a member of the National Academy of Sciences and Institute of Medicine of the National Academies, and a Fellow of the American Statistical Association. Front Matter....Pages I-XIV Introduction....Pages 1-8 A Practical Guide to Tree Construction....Pages 9-22 Logistic Regression....Pages 23-29 Classification Trees for a Binary Response....Pages 31-62 Examples Using Tree-Based Analysis....Pages 63-77 Random and Deterministic Forests....Pages 79-95 Analysis of Censored Data: Examples....Pages 97-103 Analysis of Censored Data: Concepts and Classical Methods....Pages 105-118 Analysis of Censored Data: Survival Trees and Random Forests....Pages 119-131 Regression Trees and Adaptive Splines for a Continuous Response....Pages 133-162 Analysis of Longitudinal Data....Pages 163-198 Analysis of Multiple Discrete Responses....Pages 199-225 Appendix....Pages 227-235 Back Matter....Pages 237-259