On nonparametric tests of positivity/monotonicity/convexity
Juditsky, Anatoli ; Nemirovski, Arkadi
Ann. Statist., Tome 30 (2002) no. 1, p. 498-527 / Harvested from Project Euclid
We consider the problem of estimating the distance from an unknown signal, observed in a white-noise model, to convex cones of positive/monotone/convex functions. We show that, when the unknown function belongs to a Hölder class, the risk of estimating the $L_r$-distance, $1 \leq r < \infty$, from the signal to a cone is essentially the same (up to a logarithmic factor) as that of estimating the signal itself. The same risk bounds hold for the test of positivity, monotonicity and convexity of the unknown signal. ¶ We also provide an estimate for the distance to the cone of positive functions for which risk is, by a logarithmic factor, smaller than that of the “plug-in” estimate.
Publié le : 2002-04-14
Classification:  Tests of convexity,  nonparametric test,  estimation of nonsmooth functionals,  minimax risk,  62G10,  62G08,  90C25
@article{1021379863,
     author = {Juditsky, Anatoli and Nemirovski, Arkadi},
     title = {On nonparametric tests of
			 positivity/monotonicity/convexity},
     journal = {Ann. Statist.},
     volume = {30},
     number = {1},
     year = {2002},
     pages = { 498-527},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1021379863}
}
Juditsky, Anatoli; Nemirovski, Arkadi. On nonparametric tests of
			 positivity/monotonicity/convexity. Ann. Statist., Tome 30 (2002) no. 1, pp.  498-527. http://gdmltest.u-ga.fr/item/1021379863/