No label defined (Q4344487)

This is the second of a three-volume monograph on limit theorems related with weakly dependent sequences. This volume is mainly devoted to the limiting properties of \(U\)-statistics generated by some mixing sequences. Seven chapters are included in this volume. In Chapter 1, basic concepts of \(m\)-dependent \(*\)- and \(\varphi\)-mixing are briefly introduced. Then most of the useful probability inequalities on mixing sequences are presented and some basic limit theorems on partial sums of mixing sequences are introduced, e.g., the central limit theorems, weak invariance principle, convergence rates to normal and Edgeworth expansions, etc. Chapter 2 considers the properties of functions of mixing sequences and limiting behavior of sums of functions of mixing sequences, especially the strong invariance principle. As tools and important results, some limit theorems for random variables taking values in a separable Banach space are cited. Chapters 3 and 4 are contributed to the main topic of the book. General concepts of \(U\) and \(V\) statistics generated by mixing sequences, and basic probability inequalities are introduced in Chapter 3. Chapter 4 especially deals with the limiting behavior of \(U\)-statistic of order 2, such as the Berry-Esseen type inequality, Cramér type large deviations, asymptotic expansions, etc. Chapter 5 considers the limiting behavior of the so-called generalized quadratic forms -- the \(U\)-statistics of order two whose constructing kernel function may depend on the sample size \(n\) as well as on the ordinal number of the samples. Limiting theorems presented include the complete convergence, central limit theorems and convergence to Poisson in distribution. Limiting distributions of degenerate \(U\)- and \(V\)-statistics are also considered. In Chapter 6, the basic sequence is assumed to be a normalized stationary Gaussian process. Quadratic forms can be considered as an important special case of general \(U\)-statistics. As an application of the general theory to this case, limiting behaviors are considered, some in deeper sense than those for the general \(U\) statistics, such as the iterated logarithm law. Some further applications are presented in Chapter 7, such as von Mises expansions and some other statistical expansions and asymptotic behavior of jackknife estimators.