Robust Statistics


Book Description

The Wiley-Interscience Paperback Series consists of selectedbooks that have been made more accessible to consumers in an effortto increase global appeal and general circulation. With these newunabridged softcover volumes, Wiley hopes to extend the lives ofthese works by making them available to future generations ofstatisticians, mathematicians, and scientists. "This is a nice book containing a wealth of information, much ofit due to the authors. . . . If an instructor designing such acourse wanted a textbook, this book would be the best choiceavailable. . . . There are many stimulating exercises, and the bookalso contains an excellent index and an extensive list ofreferences." —Technometrics "[This] book should be read carefully by anyone who isinterested in dealing with statistical models in a realisticfashion." —American Scientist Introducing concepts, theory, and applications, RobustStatistics is accessible to a broad audience, avoidingallusions to high-powered mathematics while emphasizing ideas,heuristics, and background. The text covers the approach based onthe influence function (the effect of an outlier on an estimater,for example) and related notions such as the breakdown point. Italso treats the change-of-variance function, fundamental conceptsand results in the framework of estimation of a single parameter,and applications to estimation of covariance matrices andregression parameters.




Influence Function Approach


Book Description

Structural mechanics is the study of the effects that forces of different physical origin (mechanical, thermal, magnetic and so on) produce on elements of structures such as cables, pillars, beams, plates and shells. This text represents the first ever attempt to include in a book format a number of standard problems from structural mechanics, which are treated by means of a single mathematical approach that is novel in the field. The influence (Green's) function method constitutes the basis for this approach. The material in this volume is based on the implementation of two important notions taken from different sciences. One of them (the influence function of a point concentrated force) is brought from structural mechanics, while the other (the Green’s function of a boundary-value problem) is taken from mathematics. They are closely related to each other, and their relation represents the keystone in this text. Bringing these notions together allows us to create a single methodological approach to a variety of problems in structural mechanics, makes their analysis easier and builds up a solid foundation for some further developments in the field. In presenting the material in this text, it was presumed that the reader's background is equally solid in undergraduate mathematics and mechanics. The reader is assumed to be relatively fluent in differential and integral calculus and to possess, at the same time, workable knowledge of the fundamental principles of statics and dynamics. Each chapter contains extensive ‘end chapter exercises’ specifically developed for each chapter, with answers and comments available in the Appendix.




Interpretable Machine Learning


Book Description

This book is about making machine learning models and their decisions interpretable. After exploring the concepts of interpretability, you will learn about simple, interpretable models such as decision trees, decision rules and linear regression. Later chapters focus on general model-agnostic methods for interpreting black box models like feature importance and accumulated local effects and explaining individual predictions with Shapley values and LIME. All interpretation methods are explained in depth and discussed critically. How do they work under the hood? What are their strengths and weaknesses? How can their outputs be interpreted? This book will enable you to select and correctly apply the interpretation method that is most suitable for your machine learning project.







Influence Functions and Matrices


Book Description

""Analyzes a wide range of problem classes originating in applied mechanics, stressing the use of influence (Green's) functions in their analysis. Provides an extensive list of influence functions and matrices-several in print for the first time. Addresses areas such as fluid flow, acoustics, electromagnetism, heat transfer, and elasticity.




Statics and Influence Functions - from a Modern Perspective


Book Description

The book teaches engineers many new things about a classical topic which suddenly is again in the center of interest because of its relevance for finite element analysis, for the accuracy of computational methods. It shows that influence functions play a fundamental role in the finite element analysis of structures and practically all of linear computational mechanics. It also strives to add new and important insights into modern structural analysis and into computational mechanics by establishing the central role of influence functions for the numerical analysis and to lay a new foundation to the energy and variational principles.




Parallel Architectures, Algorithms and Programming


Book Description

This book constitutes the refereed proceedings of the 10th International Symposium on Parallel Architectures, Algorithms and Programming, PAAP 2019, held in Guangzhou, China, in December 2019. The 39 revised full papers and 8 revised short papers presented were carefully reviewed and selected from 121 submissions. The papers deal with research results and development activities in all aspects of parallel architectures, algorithms and programming techniques.




Theory-Based Data Analysis for the Social Sciences


Book Description

This book presents the elaboration model for the multivariate analysis of observational quantitative data. This model entails the systematic introduction of "third variables" to the analysis of a focal relationship between one independent and one dependent variable to ascertain whether an inference of causality is justified. Two complementary strategies are used: an exclusionary strategy that rules out alternative explanations such as spuriousness and redundancy with competing theories, and an inclusive strategy that connects the focal relationship to a network of other relationships, including the hypothesized causal mechanisms linking the focal independent variable to the focal dependent variable. The primary emphasis is on the translation of theory into a logical analytic strategy and the interpretation of results. The elaboration model is applied with case studies drawn from newly published research that serve as prototypes for aligning theory and the data analytic plan used to test it; these studies are drawn from a wide range of substantive topics in the social sciences, such as emotion management in the workplace, subjective age identification during the transition to adulthood, and the relationship between religious and paranormal beliefs. The second application of the elaboration model is in the form of original data analysis presented in two Analysis Journals that are integrated throughout the text and implement the full elaboration model. Using real data, not contrived examples, the text provides a step-by-step guide through the process of integrating theory with data analysis in order to arrive at meaningful answers to research questions.




Power and Influence


Book Description

The world is a complex place, and this complexity is an obstacle to our attempts to explain, predict, and control it. In Power and Influence, Richard Corry investigates the assumptions that are built into the reductive method of explanation—the method whereby we study the components of a complex system in relative isolation and use the information so gained to explain or predict the behaviour of the complex whole. He investigates the metaphysical presuppositions built into the reductive method, seeking to ascertain what the world must be like in order that the method could work. Corry argues that the method assumes the existence of causal powers that manifest causal influence—a relatively unrecognised ontological category, of which forces are a paradigm example. The success of the reductive method, therefore, is an argument for the existence of such causal influences. The book goes on to show that adding causal influence to our ontology gives us the resources to solve some traditional problems in the metaphysics of causal powers, laws of nature, causation, emergence, and possibly even normative ethics. What results, then, is not just an understanding of the reductive method, but an integrated metaphysical worldview that is grounded in an ontology of power and influence.




Bayesian Networks and Influence Diagrams: A Guide to Construction and Analysis


Book Description

Bayesian Networks and Influence Diagrams: A Guide to Construction and Analysis, Second Edition, provides a comprehensive guide for practitioners who wish to understand, construct, and analyze intelligent systems for decision support based on probabilistic networks. This new edition contains six new sections, in addition to fully-updated examples, tables, figures, and a revised appendix. Intended primarily for practitioners, this book does not require sophisticated mathematical skills or deep understanding of the underlying theory and methods nor does it discuss alternative technologies for reasoning under uncertainty. The theory and methods presented are illustrated through more than 140 examples, and exercises are included for the reader to check his or her level of understanding. The techniques and methods presented for knowledge elicitation, model construction and verification, modeling techniques and tricks, learning models from data, and analyses of models have all been developed and refined on the basis of numerous courses that the authors have held for practitioners worldwide.