Consistent cross-validatory model-selection for dependent data:: hv-block cross-validation
Consistent cross-validatory model-selection for dependent data:: hv-block cross-validation
复制标题
DOI：
10.1016/s0304-4076(00)00030-0
复制
发表时间：
2000-11-01
影响因子：
6.3
通讯作者：
Racine, J
中科院分区：
文献类型：
--
作者：
Racine, J
本文考虑了 Shao (1993) 最近关于留一交叉验证模型选择渐近不一致的结果对 h 块交叉验证的影响，h 块交叉验证是 Burman、Chow 和 Nolan 提出的一种用于相关数据的交叉验证方法 (1994, Journal of Time Series Analysis 13, 189-207)。结果表明，h 块交叉验证在 Shao (1993, Journal of American Statistical Association 88(422), 486-495) 意义上是不一致的，因此不是渐近最优的。提出了一种渐近最优的 h-block 方法的修改方案，称为“hv-block”交叉验证。所提出的方法对于一般平稳观测是一致的，因为当观测总数接近无穷大时，选择具有最佳预测能力的模型的概率收敛到 1。这扩展了现有结果并产生了一种新方法，其中包含留一交叉验证、留n(v)交叉验证和h块交叉验证作为特殊情况，并考虑应用。 (C) 2000 Elsevier Science S.A. 保留所有权利。 JEL分类：C5； C51; C52。
This paper considers the impact of Shao's (1993) recent results regarding the asymptotic inconsistency of model selection via leave-one-out cross-validation on h-block cross-validation, a cross-validatory method for dependent data proposed by Burman, Chow and Nolan (1994, Journal of Time Series Analysis 13, 189-207). It is shown that h-block cross-validation is inconsistent in the sense of Shao (1993, Journal of American Statistical Association 88(422), 486-495) and therefore is not asymptotically optimal. A modification of the h-block method, dubbed 'hv-block' cross-validation, is proposed which is asymptotically optimal. The proposed approach is consistent for general stationary observations in the sense that the probability of selecting the model with the best predictive ability converges to 1 as the total number of observations approaches infinity. This extends existing results and yields a new approach which contains leave-one-out cross-validation, leave-n(v)-out cross-validation, and h-block cross-validation as special cases, Applications are considered. (C) 2000 Elsevier Science S.A. All rights reserved. JEL classification: C5; C51; C52.