Since Efron's profound paper on the bootstrap, an enormous amount of effort has been spent on the development of bootstrap, jacknife, and other resampling methods. The primary goal of these computer-intensive methods has been to provide statistical tools that work in complex situations without imposing unrealistic or unverifiable assumptions about the data generating mechanism. This book sets out to lay some of the foundations for subsampling methodology and related methods.

This volume presents selected peer-reviewed contributions from The International Work-Conference on Time Series, ITISE 2015, held in Granada, Spain, July 1-3, 2015. It discusses topics in time series analysis and forecasting, advanced methods and online learning in time series, high-dimensional and complex/big data time series as well as forecasting in real problems. The International Work-Conferences on Time Series (ITISE) provide a forum for scientists, engineers, educators and students to discuss the latest ideas and implementations in the foundations, theory, models and applications in the field of time series analysis and forecasting. It focuses on interdisciplinary and multidisciplinary research encompassing the disciplines of computer science, mathematics, statistics and econometrics.

Since I started working in the area of nonlinear programming and, later on, variational inequality problems, I have frequently been surprised to find that many algorithms, however scattered in numerous journals, monographs and books, and described rather differently, are closely related to each other. This book is meant to help the reader understand and relate algorithms to each other in some intuitive fashion, and represents, in this respect, a consolidation of the field. The framework of algorithms presented in this book is called Cost Approxi mation. (The preface of the Ph.D. thesis Pat93d] explains the background to the work that lead to the thesis, and ultimately to this book.) It describes, for a given formulation of a variational inequality or nonlinear programming problem, an algorithm by means of approximating mappings and problems, a principle for the update of the iteration points, and a merit function which guides and monitors the convergence of the algorithm. One purpose of this book is to offer this framework as an intuitively appeal ing tool for describing an algorithm. One of the advantages of the framework, or any reasonable framework for that matter, is that two algorithms may be easily related and compared through its use. This framework is particular in that it covers a vast number of methods, while still being fairly detailed; the level of abstraction is in fact the same as that of the original problem statement."

This book provides a complete exposition of equidistribution and counting problems weighted by a potential function of common perpendicular geodesics in negatively curved manifolds and simplicial trees. Avoiding any compactness assumptions, the authors extend the theory of Patterson-Sullivan, Bowen-Margulis and Oh-Shah (skinning) measures to CAT(-1) spaces with potentials. The work presents a proof for the equidistribution of equidistant hypersurfaces to Gibbs measures, and the equidistribution of common perpendicular arcs between, for instance, closed geodesics. Using tools from ergodic theory (including coding by topological Markov shifts, and an appendix by Buzzi that relates weak Gibbs measures and equilibrium states for them), the authors further prove the variational principle and rate of mixing for the geodesic flow on metric and simplicial trees-again without the need for any compactness or torsionfree assumptions. In a series of applications, using the Bruhat-Tits trees over non-Archimedean local fields, the authors subsequently prove further important results: the Mertens formula and the equidistribution of Farey fractions in function fields, the equidistribution of quadratic irrationals over function fields in their completions, and asymptotic counting results of the representations by quadratic norm forms. One of the book's main benefits is that the authors provide explicit error terms throughout. Given its scope, it will be of interest to graduate students and researchers in a wide range of fields, for instance ergodic theory, dynamical systems, geometric group theory, discrete subgroups of locally compact groups, and the arithmetic of function fields.

Statistical Analysis of Cost-effectiveness Data

Cost-effectiveness analysis is the simultaneous Statistical comparison of two or more groups with respect to costs and effectiveness. A prime example of this is the comparison of health-care interventions, where there is a growing expectation from policymakers that evidence supporting the cost-effectiveness of new interventions be provided along with customary data on efficacy and safety. Statistical Analysis of Cost-effectiveness Data provides an overview of the statistical methods used in such analysis and gives an illustrated summary of the key developments in statistical issues related to cost-effectiveness comparisons, over the last decade. Provides an up-to-date overview of statistical methods used in the analysis of cost-effectiveness data. Discusses all the major issues in the field, including: Parameter estimation for both censored and uncensored data; Making inference using cost-effectiveness ratios; Incremental net benefit plots and cost-effectiveness ratios; Incremental net benefit plots and cost-effectiveness acceptability curves; Covariate adjustment and sub-group analyses; Multinational trials; Sample size determinations using both classical and Bayesian approaches. Illustrated throughout by worked examples from the authors' own experiences.

Statistical Analysis of Cost-effectiveness Data is ideal for biostatisticians and health economists both in academia and industry. There is also much of use for graduate students of biostatistics, public health and economics, as well as individuals working in government regulator agencies.

STATISTICS IN PRACTICE

A series of practical books outlining the use of statisticaltechniques in a wide range of applications areas: HUMAN AND BIOLOGICAL SCIENCES EARTH AND ENVIRONMENTAL SCIENCES INDUSTRY, COMMERCE AND FINANCE

Categorical data arise often in many fields, including biometrics, economics, management, manufacturing, marketing, psychology, and sociology. This book provides an introduction to the analysis of such data. The coverage is broad, using the loglinear Poisson regression model and logistic binomial regression models as the primary engines for methodology. Topics covered include count regression models, such as Poisson, negative binomial, zero-inflated, and zero-truncated models; loglinear models for two-dimensional and multidimensional contingency tables, including for square tables and tables with ordered categories; and regression models for two-category (binary) and multiple-category target variables, such as logistic and proportional odds models. All methods are illustrated with analyses of real data examples, many from recent subject area journal articles. These analyses are highlighted in the text, and are more detailed than is typical, providing discussion of the context and background of the problem, model checking, and scientific implications. More than 200 exercises are provided, many also based on recent subject area literature. Data sets and computer code are available at a web site devoted to the text. Jeffrey S. Simonoff is Professor of Statistics at New York University. He is author of Smoothing Methods in Statistics and coauthor of A Casebook for a First Course in Statistics and Data Analysis, as well as numerous articles in scholarly journals. He is a Fellow of the American Statistical Association and the Institute of Mathematical Statistics, and an Elected Member of the International Statistical Institute.

Intended for graduates and researchers in physics, chemistry, biology, and applied mathematics, this book provides an up-to-date introduction to current research in fluctuations in spatially extended systems. It covers the theory of stochastic partial differential equations and gives an overview of the effects of external noise on dynamical systems with spatial degrees of freedom. Starting with a general introduction to noise-induced phenomena in dynamical systems, the text moves on to an extensive discussion of analytical and numerical tools needed to gain information from stochastic partial differential equations. It then turns to particular problems described by stochastic PDEs, covering a wide part of the rich phenomenology of spatially extended systems, such as nonequilibrium phase transitions, domain growth, pattern formation, and front propagation. The only prerequisite is a minimal background knowledge of the Langevin and Fokker-Planck equations.

This is the first book where mathematics and computer science are directly confronted and joined to tackle intricate problems in computer science with deep mathematical approaches. It contains a collection of refereed papers presented at the Colloquium on Mathematics and Computer Science held at the University of Versailles-St-Quentin on September 18-20, 2000. The colloquium was a meeting place for researchers in mathematics and computer science and thus an important opportunity to exchange ideas and points of view, and to present new approaches and new results in the common areas such as algorithms analysis, trees, combinatorics, optimization, performance evaluation and probabilities. The book is intended for a large public in applied mathematics, discrete mathematics and computer science, including researchers, teachers, graduate students and engineers. It provides an overview of the current questions in computer science and related modern mathematical methods. The range of applications is very wide and reaches beyond computer science.

This volume contains the text of the twenty-five papers presented at two workshops entitled Maximum-Entropy and Bayesian Methods in Applied Statistics, which were held at the University of Wyoming from June 8 to 10, 1981, and from August 9 to 11, 1982. The workshops were organized to bring together researchers from different fields to critically examine maxi mum-entropy and Bayesian methods in science, engineering, medicine, oceanography, economics, and other disciplines. An effort was made to maintain an informal environment where ideas could be easily ~xchanged. That the workshops were at least partially successful is borne out by the fact that there have been two succeeding workshops, and the upcoming Fifth Workshop promises to be the largest of all. These workshops and their proceedings could not have been brought to their final form without the substantial help of a number of people. The support of David Hofmann, the past chairman, and Glen Rebka, Jr. , the present chairman of the Physics Department of the University of Wyoming, has been strong and essential. Glen has taken a special interest in seeing that the proceedings have received the support required for their comple tion. The financial support of the Office of University Research Funds, University of Wyoming, is gratefully acknowledged. The secretarial staff, in particular Evelyn Haskell, Janice Gasaway, and Marce Mitchum, of the University of Wyoming Physics Department has contributed a great number of hours in helping C. Ray Smith organize and direct the workshops.

01/07 This title is now available from Walter de Gruyter. Please see www.degruyter.com for more information. This book is devoted to stochastic operators in Hilbert space. A number of models in modern probability theory apply the notion of a stochastic operator in explicit or latent form. In this book, objects from the Gaussian case are considered. Therefore, it is useful to consider all random variables and elements as functionals from the Wiener process or its formal derivative, i.e. white noise. The book consists of five chapters. The first chapter is devoted to stochastic calculus and its main goal is to prepare the tools for solving stochastic equations. In the second chapter the structure of stochastic equations, mainly the structure of Gaussian strong linear operators, is studied. In chapter 3 the definition of the action of the stochastic operator on random elements in considered. Chapter 4 deals with the mathematical models in which the notions of stochastic calculus arise and in the final chapter the equation with random operators is considered.

The author has attempted to present a book that provides a non-technical introduction into the area of non-parametric density and regression function estimation. The application of these methods is discussed in terms of the S computing environment. Smoothing in high dimensions faces the problem of data sparseness. A principal feature of smoothing, the averaging of data points in a prescribed neighborhood, is not really practicable in dimensions greater than three if we have just one hundred data points. Additive models provide a way out of this dilemma; but, for their interactiveness and recursiveness, they require highly effective algorithms. For this purpose, the method of WARPing (Weighted Averaging using Rounded Points) is described in great detail.

The dynamics of population systems cannot be understood within the framework of ordinary differential equations, which assume that the number of interacting agents is infinite. With recent advances in ecology, biochemistry and genetics it is becoming increasingly clear that real systems are in fact subject to a great deal of noise. Relevant examples include social insects competing for resources, molecules undergoing chemical reactions in a cell and a pool of genomes subject to evolution.When the population size is small, novel macroscopic phenomena can arise, which can be analyzed using the theory of stochastic processes. This thesis is centered on two unsolved problems in population dynamics: the symmetry breaking observed in foraging populations and the robustness of spatial patterns. We argue that these problems can be resolved with the help of two novel concepts: noise-induced bistable states and stochastic patterns.
"

Stochastic processes are mathematical models of random phenomena that evolve according to prescribed dynamics. Processes commonly used in applications are Markov chains in discrete and continuous time, renewal and regenerative processes, Poisson processes, and Brownian motion. This volume gives an in-depth description of the structure and basic properties of these stochastic processes. A main focus is on equilibrium distributions, strong laws of large numbers, and ordinary and functional central limit theorems for cost and performance parameters. Although these results differ for various processes, they have a common trait of being limit theorems for processes with regenerative increments. Extensive examples and exercises show how to formulate stochastic models of systems as functions of a system s data and dynamics, and how to represent and analyze cost and performance measures. Topics include stochastic networks, spatial and space-time Poisson processes, queueing, reversible processes, simulation, Brownian approximations, and varied Markovian models.

The technical level of the volume is between that of introductory texts that focus on highlights of applied stochastic processes, and advanced texts that focus on theoretical aspects of processes."

This volume presents the results of biological and medical research with the statistical methods used to obtain them. Nowadays the fields of biology and experimental medicine rely on techniques for processing of experimental data and for the evaluation of hypotheses. It is increasingly necessary to stimulate awareness of the importance of statistical techniques (and of the possible traps that they can hide) by using real data in concrete situations drawn from research activity.

An up-to-date approach to understanding statistical inference

Statistical inference is finding useful applications in numerous fields, from sociology and econometrics to biostatistics. This volume enables professionals in these and related fields to master the concepts of statistical inference under inequality constraints and to apply the theory to problems in a variety of areas.

Constrained Statistical Inference: Order, Inequality, and Shape Constraints provides a unified and up-to-date treatment of the methodology. It clearly illustrates concepts with practical examples from a variety of fields, focusing on sociology, econometrics, and biostatistics.

The authors also discuss a broad range of other inequality-constrained inference problems that do not fit well in the contemplated unified framework, providing a meaningful way for readers to comprehend methodological resolutions.

Chapter coverage includes:

• Population means and isotonic regression
• Inequality-constrained tests on normal means
• Tests in general parametric models
• Likelihood and alternatives
• Analysis of categorical data
• Inference on monotone density function, unimodal density function, shape constraints, and DMRL functions
• Bayesian perspectives, including Stein's Paradox, shrinkage estimation, and decision theory

Leading biostatisticians and biomedical researchers describe many of the key techniques used to solve commonly occurring data analytic problems in molecular biology, and demonstrate how these methods can be used in the development of new markers for exposure to a risk factor or for disease outcomes. Major areas of application include microarray analysis, proteomic studies, image quantitation, genetic susceptibility and association, evaluation of new biomarkers, and power analysis and sample size.

Lazar Mayants is a recent Russian emigre noted for his work in theoretical physics. He was previously a professor at several universities of the Soviet Union and a distinguished member of the Academy of Sciences of the U.S.S.R, where he worked for about 30 years. In this book he presents a unique, extremely detailed, and embracive version of a subject that has suffered for a long time from numerous internal imperfections. His approach is new and original, the material covered features not only the foundations of the science of probability but also most of its applications, including statistical and quantum mechanics. The key methodolOgical principle underlying the book is of extraordinary significance and deserves special attention. The treatment excels in thoroughness of presentation, in its fulness of mathe matical detail and the abundance of physical examples. The book is intended for a wide range of people interested in probability and its connection with modern science. It is written as a text for advanced students, and I predict that a reader who masters all its contents will become an expert in the subject of both prob ability and its physical implications, while enjoying its understanding and use. HENRY MARGENAU Veritas nihil veretur nisi abscondi (truth 'What tremendously easy riddles you ask ' Humpty Dumpty growled out. fears nothing except being hidden). Latin proverb Lewis Carroll, Through the Looking Glass, Chap. 6. Preface The history of producing this book is rather complicated and not quite usual."

This low-priced write-in workbook offers extensive practice, mostly context-free, for students to gain confidence in statistics at Level 3.

Introduction to the basic concepts of probability theory: independence, expectation, convergence in law and almost-sure convergence. Short expositions of more advanced topics such as Markov Chains, Stochastic Processes, Bayesian Decision Theory and Information Theory.

Michael Mitchell's A Visual Guide to Stata Graphics, Fourth Edition provides an essential introduction and reference for Stata graphics. The fourth edition retains the features that made the first three editions so useful: A complete guide to Stata's graph command Exhaustive examples of customized graphs Visual indexing of features-just look for a picture that matches what you want to do This edition includes new discussions of color, Unicode characters, export formats, sizing of graph elements, and schemes. The section on colors has been greatly expanded to include over 50 examples that demonstrate how to modify colors, add transparency, and change intensity. In the discussion of text modifications, Mitchell now shows how to include Unicode characters such as Greek letters, symbols, and emojis. New examples have also been added that show how to change the size of graph elements such as text, markers, and line widths using both absolute units (points, inches, and centimeters) as well as relative units (line large or *2 for two times the original size). Finally, the look of graphs throughout the book has changed-most graphs are now created using a common updated scheme. The book's visual style makes it easy to find exactly what you need. A color-coded, visual table of contents runs along the edge of every page and shows readers exactly where they are in the book. You can see the color-coded chapter tabs without opening the book, providing quick visual access to each chapter. The heart of each chapter is a series of entries that are typically formatted three to a page. Each entry shows a graph command (with the emphasized portion of the command highlighted in red), the resulting graph, a description of what is being done, and the dataset used. Because every feature, option, and edit is demonstrated with a graph, you can often flip through a section of the book to find exactly the effect you are seeking. The book begins with an introduction to Stata graphs that includes an overview of graphs types, schemes, and options and the process of building a graph. Then, it turns to detailed discussions of many graph types-scatterplots, regression fit plots, line plots, contour plots, bar graphs, box plots, and many others. Mitchell shows how to create each type of graph and how to use options to control the look of the graph. Because Stata's graph command will let you customize any aspect of the graph, Mitchell spends ample time showing you the most valuable options for obtaining the look you want. If you are in a hurry to discover one special option, you can skim the chapter until you see the effect you want and then glance at the command to see what is highlighted in red. After focusing on specific types of graphs, Mitchell undertakes an in-depth presentation of the options available across almost all graph types. This includes options that add and change the look of titles, notes, and such; control the number of ticks on axes; control the content and appearance of the numbers and labels on axes; control legends; add and change the look of annotations; graph over subgroups; change the look of markers and their labels; size graphs and their elements; and more. To complete the graphical journey, Mitchell discusses and demonstrates the 12 styles that unite and control the appearance of the myriad graph objects. These styles are angles, colors, clock positions, compass directions, connecting points, line patterns, line widths, margins, marker sizes, orientations, marker symbols, and text sizes. You won't want to overlook the appendix in this book. There Mitchell first gives a quick overview of the dozens of statistical graph commands that are not strictly the subject of the book. Even so, these commands use the graph command as an engine to draw their graphs; therefore, almost all that Mitchell has discussed applies to them. He also addresses combining graphs-showing you how to create complex and multipart images from previously created graphs. In a crucial section titled "Putting it all together", Mitchell shows us how to do just that. We learn more about overlaying twoway plots, and we learn how to combine data management and graphics to create plots such as bar charts of rates with capped confidence intervals. Mitchell concludes by warning us about mistakes that can be made when typing graph commands and how to correct them. The fourth edition of A Visual Guide to Stata Graphics is a complete guide to Stata's graph command and the associated Graph Editor. Whether you want to tame the Stata graph command, quickly find out how to produce a graphical effect, or learn approaches that can be used to construct custom graphs, this is the book to read.

This textbook provides a comprehensive introduction to probability and stochastic processes, and shows how these subjects may be applied in computer performance modeling. The author's aim is to derive probability theory in a way that highlights the complementary nature of its formal, intuitive, and applicative aspects while illustrating how the theory is applied in a variety of settings. Readers are assumed to be familiar with elementary linear algebra and calculus, including being conversant with limits, but otherwise, this book provides a self-contained approach suitable for graduate or advanced undergraduate students. The first half of the book covers the basic concepts of probability, including combinatorics, expectation, random variables, and fundamental theorems. In the second half of the book, the reader is introduced to stochastic processes. Subjects covered include renewal processes, queueing theory, Markov processes, matrix geometric techniques, reversibility, and networks of queues. Examples and applications are drawn from problems in computer performance modeling. Throughout, large numbers of exercises of varying degrees of difficulty will help to secure a reader's understanding of these important and fascinating subjects.

This book introduces the reader to the use of Monte Carlo methods for solving practical problems in radiation transport, and will also serve as a reference work for practitioners in the field. It assumes the reader has a general knowledge of calculus and radiation physics, and a knowledge of Fortran programming, but assumes no prior knowledge of stochastic methods or statistical physics. The subject is presented by a combination of theoretical development and practical calculations. Because Monte Carlo methods are closely linked to the use of computers, from the beginning the reader is taught to convert the theoretical constructs developed in the text into functional software for use on a personal computer. Example problems provide the reader with an in-depth understanding of the concepts presented and lead to the production of a unique learning tool, a probabilistic framework code that models in a simple manner the features of production of Monte Carlo transport codes. This framework code is developed in stages such that every function is understood, tested, and demonstrated - random sampling, generating random numbers, implementing geometric models, using variance reduction, tracking particles in a random walk, testing the thoroughness with which the problem phase space is sampled, scoring detectors, and obtaining estimates of uncertainty in results. Advanced topics covered include criticality, correlated sampling, adjoint transport, and neutron thermalization. Monte Carlo codes can produce highly precise wrong answers. The probability of this occurring is increased if production codes are run as opaque, black boxes' of software. This text attempts to make Monte Carlo into acomprehensible, usable tool for solving practical transport problems. It is suitable for advanced undergraduate and graduate students and researchers who wish to expand their knowledge of the Monte Carlo technique.