Almost sure central limit theorems for functionals of absolutely regular processes with application to \(U\)-statistics (Q2470504): Difference between revisions

Given a stationary sequence of random variables \(\{ X_n \}\) and a symmetric function \(h: \mathbb{R}^2 \rightarrow \mathbb{R}\), the \(U\)-statistic with kernel \(h\) is defined by \(U_n = {n \choose 2}^{-1} \sum_{1\leq i<j\leq n} h(X_i, X_j).\) Assume \(E U_n= \theta.\) An almost sure central limit theorem holds for the \(U\)-statistic if \[ \lim_{n\rightarrow \infty} (\log n)^{-1} \sum_{k=1}^{n} k^{-1} I( \sqrt{k}(U_k - \theta) \leq 2 \sigma x) = \Phi(x) \;\text{a.s.}, \] where \(\Phi(x)\) denotes the distribution function for the standard normal random variable and \(\sigma\) is a constant. The paper develops an almost sure central limit theorem when \(\{ X_n \}\) is a 1-approximating functional of an absolutely regular process and then extends it to obtain the corresponding result for \(U\)-statistics based on such processes.