Non-parametric estimation for partially observed transient diffusion processes
作者:
V. Genon-Catalot,
C. Laredo,
期刊:
Stochastics
(Taylor Available online 1986)
卷期:
Volume 18,
issue 2
页码: 169-196
ISSN:0090-9491
年代: 1986
DOI:10.1080/17442508608833407
出版商: Gordon and Breach, Science Publishers, Inc
关键词: Diffusion processes;hitting times;processes with independent increments;Girsanov formula;non-parametric estimation;point processes
数据来源: Taylor
摘要:
Consider Xta transient diffusion on R with small known variance σ and unknown drift function µ(.). The record processis partially observed up to the hitting time T$sub:ⱥ$esub: of a given level ⱥ>x = X$sub:0$esub: Under suitable assumptions on µ, we study the statistical model corresponding to this observation. The likelihood function being in most cases unknown, we construct a non-parametric estimator of the drift coefficient pi µ(.) and study its asymptotic properties as σgoes to 0.
点击下载:
PDF (658KB)
返 回