Abstract:
We study the asymptotic behavior of the minimax risk in nonparametric regression estimation in the uniform norm in classes of functions, satisfying the Lipschitz or Holder condition. For an arbitrary loss function, the exact asymptotic behavior of the risk under equidistant observation design and Gaussian noise is found.
Citation:
A. P. Korostelev, “An asymptotically minimax regression estimator in the uniform norm up to exact constant”, Teor. Veroyatnost. i Primenen., 38:4 (1993), 875–882; Theory Probab. Appl., 38:4 (1993), 737–743