Approximate leave-future-out cross-validation for Bayesian time series models

被引:54
|
作者
Burkner, Paul-Christian [1 ]
Gabry, Jonah [2 ,3 ]
Vehtari, Aki [1 ]
机构
[1] Aalto Univ, Dept Comp Sci, Konemiehentie 2, Espoo 02150, Finland
[2] Columbia Univ, Appl Stat Ctr, New York, NY USA
[3] Columbia Univ, ISERP, New York, NY USA
基金
芬兰科学院;
关键词
Time series analysis; cross-Validation; Bayesian inference; pareto Smoothed importance sampling; R PACKAGE;
D O I
10.1080/00949655.2020.1783262
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
One of the common goals of time series analysis is to use the observed series to inform predictions for future observations. In the absence of any actual new data to predict, cross-validation can be used to estimate a model's future predictive accuracy, for instance, for the purpose of model comparison or selection. Exact cross-validation for Bayesian models is often computationally expensive, but approximate cross-validation methods have been developed, most notably methods for leave-one-out cross-validation (LOO-CV). If the actual prediction task is to predict the future given the past, LOO-CV provides an overly optimistic estimate because the information from future observations is available to influence predictions of the past. To properly account for the time series structure, we can use leave-future-out cross-validation (LFO-CV). Like exact LOO-CV, exact LFO-CV requires refitting the model many times to different subsets of the data. Using Pareto smoothed importance sampling, we propose a method for approximating exact LFO-CV that drastically reduces the computational costs while also providing informative diagnostics about the quality of the approximation.
引用
收藏
页码:2499 / 2523
页数:25
相关论文
共 50 条
  • [31] Using cross-validation methods to select time series models: Promises and pitfalls
    Liu, Siwei
    Zhou, Di Jody
    BRITISH JOURNAL OF MATHEMATICAL & STATISTICAL PSYCHOLOGY, 2024, 77 (02): : 337 - 355
  • [32] Unbiased estimator for the variance of the leave-one-out cross-validation estimator for a Bayesian normal model with fixed variance
    Sivula, Tuomas
    Magnusson, Mans
    Vehtari, Aki
    COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 2023, 52 (16) : 5877 - 5899
  • [33] A leave-K-out cross-validation scheme for unsupervised kernel regression
    Klanke, Stefan
    Ritter, Helge
    ARTIFICIAL NEURAL NETWORKS - ICANN 2006, PT 2, 2006, 4132 : 427 - 436
  • [34] Quicksort leave-pair-out cross-validation for ROC curve analysis
    Numminen, Riikka
    Montoya Perez, Ileana
    Jambor, Ivan
    Pahikkala, Tapio
    Airola, Antti
    COMPUTATIONAL STATISTICS, 2023, 38 (03) : 1579 - 1595
  • [35] ASYMPTOTIC OPTIMALITY AND EFFICIENT COMPUTATION OF THE LEAVE-SUBJECT-OUT CROSS-VALIDATION
    Xu, Ganggang
    Huang, Jianhua Z.
    ANNALS OF STATISTICS, 2012, 40 (06): : 3003 - 3030
  • [36] Nonparametric density estimation by exact leave-p-out cross-validation
    Celisse, Alain
    Robin, Stephane
    COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2008, 52 (05) : 2350 - 2368
  • [37] Quicksort leave-pair-out cross-validation for ROC curve analysis
    Riikka Numminen
    Ileana Montoya Perez
    Ivan Jambor
    Tapio Pahikkala
    Antti Airola
    Computational Statistics, 2023, 38 : 1579 - 1595
  • [38] Nearest neighbour distance matching Leave-One-Out Cross-Validation for map validation
    Mila, Carles
    Mateu, Jorge
    Pebesma, Edzer
    Meyer, Hanna
    METHODS IN ECOLOGY AND EVOLUTION, 2022, 13 (06): : 1304 - 1316
  • [39] Leave-one-ion-out cross-validation for assisting in developing robust QSPR models of ionic liquids
    Liu, Xiao
    Yu, Mengxian
    Jia, Qingzhu
    Yan, Fangyou
    Zhou, Yin-Ning
    Wang, Qiang
    JOURNAL OF MOLECULAR LIQUIDS, 2023, 388
  • [40] Learning dynamic Bayesian network models via cross-validation
    Peña, JM
    Björkegren, J
    Tegnér, J
    PATTERN RECOGNITION LETTERS, 2005, 26 (14) : 2295 - 2308