$U$-Statistic Processes: A Martingale Approach
Stute, Winfried
Ann. Probab., Tome 22 (1994) no. 4, p. 1725-1744 / Harvested from Project Euclid
For i.i.d. data $X_1,\cdots, X_n$ and a kernel $h$, the associated $U$-statistic process is defined as $U_n (u, v) = \frac{1}{n(n-1)} \sum_{1\leq i\neq j\leq n} h(X_i, X_j)1_{\{X_i\leq u,X_j\leq \nu\}}.$ Variants of these processes occur, for example, in the representation of the product-limit estimator of a lifetime distribution for censored/truncated data or in trimmed $U$-statistics. We derive an almost sure representation of $U_n$ under weak moment assumptions on $h$. Proofs rely on a proper decomposition of the remainder term into strong two-parameter martingales.
Publié le : 1994-10-14
Classification:  $U$-statistic processes,  maximal inequalities,  martingale decomposition,  Hajek projection,  62G30,  60G42,  62G05
@article{1176988480,
     author = {Stute, Winfried},
     title = {$U$-Statistic Processes: A Martingale Approach},
     journal = {Ann. Probab.},
     volume = {22},
     number = {4},
     year = {1994},
     pages = { 1725-1744},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1176988480}
}
Stute, Winfried. $U$-Statistic Processes: A Martingale Approach. Ann. Probab., Tome 22 (1994) no. 4, pp.  1725-1744. http://gdmltest.u-ga.fr/item/1176988480/