This research-level reference gives a unified example of U-statistic theory, with applications to indicate the range and scope of U-statistic methods-providing a concise, up-to-date source of theories accumulated over the past 40 years. Presenting numerous examples to clarify important concepts, U-Statistics describes the optimal properties of U-statistics ... explores variations of the basic theory ... explains asymptotic theory ... studies related classes of statistics ... treats the problem of estimating standard errors of U-statistics ... and suggests applications of U-statistics in statistical practice. Complete with over 200 end-of-chapter references, U-Statistics is an invaluable addition to the libraries of applied and theoretical statisticians and mathematicians, numerical analysts, and graduate students in statistics. Book jacket.

A timely and applied approach to the newly discovered methods and applications of U-statistics Built on years of collaborative research and academic experience, Modern Applied U-Statistics successfully presents a thorough introduction to the theory of U-statistics using in-depth examples and applications that address contemporary areas of study including biomedical and psychosocial research. Utilizing a "learn by example" approach, this book provides an accessible, yet in-depth, treatment of U-statistics, as well as addresses key concepts in asymptotic theory by integrating translational and cross-disciplinary research. The authors begin with an introduction of the essential and theoretical foundations of U-statistics such as the notion of convergence in probability and distribution, basic convergence results, stochastic Os, inference theory, generalized estimating equations, as well as the definition and asymptotic properties of U-statistics. With an emphasis on nonparametric applications when and where applicable, the authors then build upon this established foundation in order to equip readers with the knowledge needed to understand the modern-day extensions of U-statistics that are explored in subsequent chapters. Additional topical coverage includes: Longitudinal data modeling with missing data Parametric and distribution-free mixed-effect and structural equation models A new multi-response based regression framework for non-parametric statistics such as the product moment correlation, Kendall's tau, and Mann-Whitney-Wilcoxon rank tests A new class of U-statistic-based estimating equations (UBEE) for dependent responses Motivating examples, in-depth illustrations of statistical and model-building concepts, and an extensive discussion of longitudinal study designs strengthen the real-world utility and comprehension of this book. An accompanying Web site features SAS? and S-Plus? program codes, software applications, and additional study data. Modern Applied U-Statistics accommodates second- and third-year students of biostatistics at the graduate level and also serves as an excellent self-study for practitioners in the fields of bioinformatics and psychosocial research.

The theory of U-statistics goes back to the fundamental work of Hoeffding [1], in which he proved the central limit theorem. During last forty years the interest to this class of random variables has been permanently increasing, and thus, the new intensively developing branch of probability theory has been formed. The U-statistics are one of the universal objects of the modem probability theory of summation. On the one hand, they are more complicated "algebraically" than sums of independent random variables and vectors, and on the other hand, they contain essential elements of dependence which display themselves in the martingale properties. In addition, the U -statistics as an object of mathematical statistics occupy one of the central places in statistical problems. The development of the theory of U-statistics is stipulated by the influence of the classical theory of summation of independent random variables: The law of large num bers, central limit theorem, invariance principle, and the law of the iterated logarithm we re proved, the estimates of convergence rate were obtained, etc.

Conventional statistical methods routinely miss differences among groups or associations among variables. These differences are detected by more modern techniques. Hundreds of journal articles have described the reasons why standard techniques are unsatisfactory. Nonetheless, simple and intuitive explanations are generally unavailable. Without assuming any prior training in statistics, Part I of this book describes basic statistical principles from a point of view that makes their shortcomings easy to understand. Part II describes modern methods that address the problems covered in Part I. Using data from actual studies, many examples are included.

Samples and surveys; The analysis of variance; Nonparametric statistical tests of hypotheses; Rank estimates from nonparametric tests; Statistical inferences for ordered parameters: a personal view of isotonic regression since the work by Barlow, Bartholomew, Bremner and Brunk; Time series: model estimation, data analysis and robust procedures.

This is an introductory text on a broad class of statistical estimators that are minimizers of convex functions. It covers the basics of U-statistics and Mm-estimators and develops their asymptotic properties. It also provides an elementary introduction to resampling, particularly in the context of these estimators. The last chapter is on practical implementation of the methods presented in other chapters, using the free software R.