Slow stochastic approximation processes are good for estimating slope
Teoriâ veroâtnostej i ee primeneniâ, Tome 40 (1995) no. 1, pp. 192-199
Cet article a éte moissonné depuis la source Math-Net.Ru
A continuous-time Robbins–Monroe process violating the conditions necessary for the CLT to hold will be considered. It will be shown that although the estimator process 6% converges to the root to be determined a.s. it is sufficiently rich to get strong consistent estimator of the slope of the regressor function using noisy observations of the regressor function at $\theta_t-s$ only.
Keywords:
stochastic approximation, least square estimation, stochastic regression, the Lai–Wei condition, the Cameron–Martin formula.
@article{TVP_1995_40_1_a15,
author = {L. Gerencs\'er},
title = {Slow stochastic approximation processes are good for estimating slope},
journal = {Teori\^a vero\^atnostej i ee primeneni\^a},
pages = {192--199},
year = {1995},
volume = {40},
number = {1},
language = {en},
url = {http://geodesic.mathdoc.fr/item/TVP_1995_40_1_a15/}
}
L. Gerencsér. Slow stochastic approximation processes are good for estimating slope. Teoriâ veroâtnostej i ee primeneniâ, Tome 40 (1995) no. 1, pp. 192-199. http://geodesic.mathdoc.fr/item/TVP_1995_40_1_a15/