On bandwidth choice in nonparametric regression with both short- and long-range dependent errors
Hall, Peter ; Lahiri, Soumendra Nath ; Polzehl, Jörg
Ann. Statist., Tome 23 (1995) no. 6, p. 1921-1936 / Harvested from Project Euclid
We analyse methods based on the block bootstrap and leave-out cross-validation, for choosing the bandwidth in nonparametric regression when errors have an almost arbitrarily long range of dependence. A novel analytical device for modelling the dependence structure of errors is introduced. This allows a concise theoretical description of the way in which the range of dependence affects optimal bandwidth choice. It is shown that, provided block length or leave-out number, respectively, are chosen appropriately, both techniques produce first-order optimal bandwidths. Nevertheless, the block bootstrap has far better empirical properties, particularly under long-range dependence.
Publié le : 1995-12-14
Classification:  Bandwidth choice,  block bootstrap,  correlated errors,  cross-validation,  curve estimation,  kernel estimator,  local linear smoothing,  long-range dependence,  mean squared error,  nonparametric regression,  resampling,  short-range dependence,  62G07,  62G09,  62M10
@article{1034713640,
     author = {Hall, Peter and Lahiri, Soumendra Nath and Polzehl, J\"org},
     title = {On bandwidth choice in nonparametric regression with both short- and long-range dependent errors},
     journal = {Ann. Statist.},
     volume = {23},
     number = {6},
     year = {1995},
     pages = { 1921-1936},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1034713640}
}
Hall, Peter; Lahiri, Soumendra Nath; Polzehl, Jörg. On bandwidth choice in nonparametric regression with both short- and long-range dependent errors. Ann. Statist., Tome 23 (1995) no. 6, pp.  1921-1936. http://gdmltest.u-ga.fr/item/1034713640/