Stationary bootstrap for kernel density estimators under -weak dependence

详细信息	查看全文 | 推荐本文 |

• 作者：Eunju Hwang ; Dong Wan Shin ; ^{shindw@ewha.ac.kr} ; ^{shindw@mm.ewha.ac.kr}
• 关键词：Stationary bootstrap ; Weak dependence ; Kernel estimator ; Density ; Derivative
• 刊名：Computational Statistics and Data Analysis
• 出版年：2012
• 期刊代码：19_01679473
• 类别：cp
• 出版时间：June, 2012
• 卷：56
• 期：6
• 页码：1581-1593
• 文件大小：368 K

摘要

Stationary bootstrap technique is applied for kernel-type estimators of densities and their derivatives of stationary -weakly dependent processes. The -weak dependence, introduced by Doukhan & Louhichi [Doukhan, P., Louhichi, S., 1999. A new weak dependence condition and applications to moment inequalities. Stochastic Processes and their Applications 84, 313-342], unifies weak dependence conditions such as mixing, association, Gaussian sequences and Bernoulli shifts. The class of -weakly dependent processes includes all weakly dependent processes of interest in statistics, containing such important processes as GARCH processes, threshold autoregressive processes, and bilinear processes. We obtain asymptotic validity for the stationary bootstrap in the density and derivatives estimation. A Monte-Carlo experiment compares the proposed method with other methods. Log returns of daily Dow Jones index are analyzed by the proposed method.