Data-driven rate-optimal specification testing in regression models
Guerre, Emmanuel ; Lavergne, Pascal
Ann. Statist., Tome 33 (2005) no. 1, p. 840-870 / Harvested from Project Euclid
We propose new data-driven smooth tests for a parametric regression function. The smoothing parameter is selected through a new criterion that favors a large smoothing parameter under the null hypothesis. The resulting test is adaptive rate-optimal and consistent against Pitman local alternatives approaching the parametric model at a rate arbitrarily close to $1/\sqrt{n}$ . Asymptotic critical values come from the standard normal distribution and the bootstrap can be used in small samples. A general formalization allows one to consider a large class of linear smoothing methods, which can be tailored for detection of additive alternatives.
Publié le : 2005-04-14
Classification:  Hypothesis testing,  nonparametric adaptive tests,  selection methods,  62G10,  62G08
@article{1117114338,
     author = {Guerre, Emmanuel and Lavergne, Pascal},
     title = {Data-driven rate-optimal specification testing in regression models},
     journal = {Ann. Statist.},
     volume = {33},
     number = {1},
     year = {2005},
     pages = { 840-870},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1117114338}
}
Guerre, Emmanuel; Lavergne, Pascal. Data-driven rate-optimal specification testing in regression models. Ann. Statist., Tome 33 (2005) no. 1, pp.  840-870. http://gdmltest.u-ga.fr/item/1117114338/