Deviance information criterion for latent variable models and misspecified models

Deviance information criterion for latent variable models and misspecified models
复制标题

潜变量模型和错误指定模型的偏差信息准则

DOI:
10.1016/j.jeconom.2019.11.002
复制
发表时间:
2020-06
影响因子:
6.3
通讯作者:
李勇
李勇
中科院分区:
经济学2区
文献类型:
--
作者:
李勇

文献摘要

参考文献

被引文献

相似文献

偏差信息准则(DIC)已广泛用于贝叶斯模型比较,特别是在使用马尔可夫链蒙特卡罗(MCMC)估计候选模型之后。本文首先研究了当根据条件似然计算DIC时使用DIC比较潜变量模型的问题。特别是,结果表明条件似然方法破坏了 DIC 的理论基础。提出了 DIC 的新版本,即 DIC L,用于比较潜变量模型。研究了DIC L的大样本特性。提供了 DIC L 的频率论理由。与 AIC 一样,DIC L 为 DGP 和预测分布之间的预期 Kullback-Leibler (KL) 散度提供了渐近无偏估计量。引入了一些流行的算法,如 EM、卡尔曼和粒子滤波算法来计算潜变量模型的 DIC L。此外,本文还研究了使用 DIC 来比较错误指定模型的问题。提出了 DIC 的新版本,即 DIC M,它可以被视为 TIC 的贝叶斯版本。 DIC M 的频率论理由是在错误指定下提供的。 DIC L 和 DIC M 使用资产定价模型和随机波动率模型进行说明。
Deviance information criterion (DIC) has been widely used for Bayesian model comparison, especially after Markov chain Monte Carlo (MCMC) is used to estimate candidate models. This paper first studies the problem of using DIC to compare latent variable models when DIC is calculated from the conditional likelihood. In particular, it is shown that the conditional likelihood approach undermines theoretical underpinnings of DIC. A new version of DIC, namely DIC L, is proposed to compare latent variable models. The large sample properties of DIC L are studied. A frequentist justification of DIC L is provided. Like AIC, DIC L provides an asymptotically unbiased estimator to the expected Kullback–Leibler (KL) divergence between the DGP and a predictive distribution. Some popular algorithms, such as the EM, Kalman and particle filtering algorithms, are introduced to compute DIC L for latent variable models. Moreover, this paper studies the problem of using DIC to compare misspecified models. A new version of DIC, namely DIC M, is proposed and it can be regarded as a Bayesian version of TIC. A frequentist justification of DIC M is provided under misspecification. DIC L and DIC M are illustrated using asset pricing models and stochastic volatility models.
DOI: 10.1016/b978-0-12-820074-2.00007-1
发表时间: 2021
期刊: Local Electricity Markets
影响因子: --
作者:
Ella Mocanu;D. Mocanu;N. Paterakis;Madeleine Gibescu
通讯作者: Ella Mocanu;D. Mocanu;N. Paterakis;Madeleine Gibescu
DOI: 10.1017/s0266466600008768
发表时间: 1994-08
期刊: Econometric Theory
影响因子: 0.8
作者:
Jae-Young Kim
通讯作者: Jae-Young Kim
DOI: 10.1080/01621459.1995.10476592
发表时间: 1995-09
影响因子: 3.7
作者:
R. Kass;L. Wasserman
通讯作者: R. Kass;L. Wasserman
DOI: 10.4324/9780203491287
发表时间: 1995
期刊: --
影响因子: --
作者:
Andrew Gelman;John B. Carlin;H. Stern;David B. Dunson;A. Vehtari;Donald B. Rubin
通讯作者: Andrew Gelman;John B. Carlin;H. Stern;David B. Dunson;A. Vehtari;Donald B. Rubin
DOI: 10.1080/01621459.1976.10480949
发表时间: 1976-12
影响因子: 3.7
作者:
G. Box
通讯作者: G. Box