No label defined (Q3979862)

Let \(X_ 1,\dots,X_ n\) be independent identically distributed random variables with values in a measurable space \(({\mathcal X},{\mathcal A})\). By \(UB\)-statistics the author means the statistics \[ U_ n={n \choose m}^{-1}\sum_{1\leq i_ 1<\dots<i_ m\leq n}\Phi(X_{i_ 1},\dots,X_{i_ m}), \] where \(\Phi: {\mathcal X}^ m\to B\) is a symmetric measurable function with values in a real separable Banach space \(B\). If \(\Phi\) is such that \(E\|\Phi(X_ 1,\dots,X_ m)\|<\infty\), then \(U_ n\) is an unbiased estimate for \(\theta=E \Phi(X_ 1,\dots,X_ m)\in B\). The author establishes inequalities for \(UB\)-statistics and investigates the speed of convergence in the law of large numbers. It is shown that if \(B\) is of type 2 and \(p\)-uniformly smooth, \(4/3\leq p\leq 2\), then for \(UB\)-statistics of rank 1 the central limit theorem holds, i.e. \((\sqrt n /m)(U_ n-\theta)\) converges in distribution to a \(B\)- valued Gaussian random element. The author also finds conditions for \(n^{r/2}(U_ n-\theta)\) to converge in distribution when \(UB\)- statistics \(U_ n\) possesses a rank \(r\geq 1\).