A decomposition for generalized U-statistics of long-memory linear processes.
The paper considers a class of statistics based on a sequence of observations, \(\{X_n\}\), from a linear process with long range dependence. The statistics have the form \[ U_N= \sum_{1\leq n_1, n_2\leq N} K_{n_1,n_2}(X_{n_1}, X_{n_2}), \] where for each \(n_1, n_2\), \(K_{n_1,n_2}\) is a symmetric function depending on the indices through \(| n_1-n_2| .\) The class includes sample moments, sample autocovariance functions and various quadratic forms. A technique is described for decomposing the statistics into uncorrelated long memory and short memory components. These decompositions then enable asymptotic results to be established. For the entire collection see [Zbl 1005.00017].
- Estimating invariant laws of linear processes by U-statistics.
- Two-sample \(U\)-statistic processes for long-range dependent data
- A decomposition for some \(U\)-type statistics
- \(U\)-processes, \(U\)-quantile processes and generalized linear statistics of dependent data
- Asymptotic properties of U-processes under long-range dependence
- scientific article; zbMATH DE number 847242
- scientific article; zbMATH DE number 1925804
- Decoupling and limit theorems for U-statistics and U-processes
- Bivariate symmetric statistics of long-range dependent observations
- Long-range dependent time series specification
- Long-range dependence and the rank of decompositions
- Memory properties of transformations of linear processes
- Asymptotics for statistical functionals of long-memory sequences
- Asymptotic theory for curve-crossing analysis
This page was built for publication: A decomposition for generalized U-statistics of long-memory linear processes.
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4407602)