Nils Lid Hjort - Böcker

Highly Structured Stochastic Systems (HSSS) is a modern strategy for building statistical models for challenging real-world problems, for computing with them, and for interpreting the resulting inferences. Complexity is handled by working up from simple local assumptions in a coherent way, and that is the key to modelling, computation, inference and interpretation; the unifying framework is that of Bayesian hierarchical models. The aim of this book is to make recent developments in HSSS accessible to a general statistical audience. Graphical modelling and Markov chain Monte Carlo (MCMC) methodology are central to the field, and in this text they are covered in depth. The chapters on graphical modelling focus on causality and its interplay with time, the role of latent variables, and on some innovative applications. Those on Monte Carlo algorithms include discussion of the impact of recent theoretical work on the evaluation of performance in MCMC, extensions to variable dimension problems, and methods for dynamic problems based on particle filters. Coverage of these underlying methodologies is balanced by substantive areas of application - in the areas of spatial statistics (with epidemiological, ecological and image analysis applications) and biology (including infectious diseases, gene mapping and evolutionary genetics). The book concludes with two topics (model criticism and Bayesian nonparametrics) that seek to challenge the parametric assumptions that otherwise underlie most HSSS models. Altogether there are 15 topics in the book, and for each there is a substantial article by a leading author in the field, and two invited commentaries that complement, extend or discuss the main article, and should be read in parallel. All authors are distinguished researchers in the field, and were active participants in an international research programme on HSSS.This is the 27th volume in the Oxford Statistical Science Series, which includes texts and monographs covering many topics of current research interest in pure and applied statistics. These texts focus on topics that have been at the forefront of research interest for several years. Other books in the series include: J.Durbin and S.J.Koopman: Time series analysis by State Space Models; Peter J. Diggle, Patrick Heagerty, Kung-Yee Liang, Scott L. Zeger: Analysis of Longitudinal Data 2/e; J.K. Lindsey: Nonlinear Models in Medical Statistics; Peter J. Green, Nils L. Hjort & Sylvia Richardson: Highly Structured Stochastic Systems; Margaret S. Pepe: Statistical Evaluation of Medical Tests.

Bayesian nonparametrics works - theoretically, computationally. The theory provides highly flexible models whose complexity grows appropriately with the amount of data. Computational issues, though challenging, are no longer intractable. All that is needed is an entry point: this intelligent book is the perfect guide to what can seem a forbidding landscape. Tutorial chapters by Ghosal, Lijoi and Prünster, Teh and Jordan, and Dunson advance from theory, to basic models and hierarchical modeling, to applications and implementation, particularly in computer science and biostatistics. These are complemented by companion chapters by the editors and Griffin and Quintana, providing additional models, examining computational issues, identifying future growth areas, and giving links to related topics. This coherent text gives ready access both to underlying principles and to state-of-the-art practice. Specific examples are drawn from information retrieval, NLP, machine vision, computational biology, biostatistics, and bioinformatics.

Given a data set, you can fit thousands of models at the push of a button, but how do you choose the best? With so many candidate models, overfitting is a real danger. Is the monkey who typed Hamlet actually a good writer? Choosing a model is central to all statistical work with data. We have seen rapid advances in model fitting and in the theoretical understanding of model selection, yet this book is the first to synthesize research and practice from this active field. Model choice criteria are explained, discussed and compared, including the AIC, BIC, DIC and FIC. The uncertainties involved with model selection are tackled, with discussions of frequentist and Bayesian methods; model averaging schemes are presented. Real-data examples are complemented by derivations providing deeper insight into the methodology, and instructive exercises build familiarity with the methods. The companion website features Data sets and R code.