Bayesian multivariate semi-Markov-switching mixed data sampling

Abstract

We introduce a flexible model for multivariate time-series exhibiting heterogeneous sampling frequencies, where time-varying unobservable heterogeneity is captured by a finite number of latent regimes. The latent unobservable process evolves over time according to a semi-Markov chain. The inference is based on a Bayesian approach involving reversible jump Markov chain Monte Carlo (RJ-MCMC), which allows us to avoid specifying the number of latent regimes in advance. In a simulation study we show how our approach correctly recovers the true configuration of latent regimes with high probability, and that the proposed model can be seen as advantageous with respect to possible competitors. We illustrate through an analysis of two major polluting agents recorded daily at the Danmarksplass site (Norway) in 2022, and their association with certain mixed-frequency weather variables.

Keywords

Discrete latent variables mixed-data sampling (MIDAS)reversible jump transdimensional Markov Chain Monte Carlo (MCMC)

Get full access to this article

View all access options for this article.

References

, Dang

, Wang

, Zhou

and Mai

(2024) Mixed-frequency data sampling Grey system model: Forecasting annual CO2 emissions in China with quarterly and monthly economic-energy indicators. Applied Energy , 370, 123531.

Andreou

, Ghysels

and Kourtellos

(2010) Regression models with mixed sampling frequencies. Journal of Econometrics , 158, 246–61.

Baccini

, Biggeri

, Lagazio

, Lertxundi

and Saez

(2007) Parametric and semiparametric approaches in the analysis of short-term effects of air pollution on health. Computational Statistics and Data Analysis , 51, 4324–36.

Bartolucci

and Pandolfi

(2018) Bayesian inference for a class of latent Markov models for categorical longitudinal data. arXiv:1101.0391v2.

Bartolucci

, Farcomeni

and Pennoni

(2013) Latent Markov Models for Longitudinal Data . Boca Raton, FL: Chapman and Hall/CRC Press.

Bonas

, Datta

, Wikle

, Boone

, Alamri

, Hari

, Kavila

, Simmons

, Jarvis

, Burr

, . (2024) Assessing predictability of environmental time series with statistical and machine learning models. Environmetrics , 36, e2864.

Bottazzi Schenone

, Grimaccia

and Vichi

(2024) Structural equation models for simultaneous modeling of air pollutants. Environmetrics , 35, e2837.

Bremer

, Johnson

, Flottum

, Kverndokk

, Wardekker

and Kraus

(2020) Portrait of a climate city: How climate change is emerging as a risk in Bergen, Norway. Climate Risk Management , 29, 100236.

Bulla

and Bulla

(2006) Stylized facts of financial time series and hidden semi-Markov models. Computational Statistics and Data Analysis , 51, 2192–2209.

10.

Chan

JCC

, Poon

and Zhu

(2024) Timevarying parameter MIDAS models: Application to nowcasting US real GDP. Technical report, SSRN: Available at: https://ssrn.com/abstract=4802887

11.

Chang

, C-M

Hsu

, S-T

Chen

, M-C

Wang

and C-F

(2023) Revisiting economic growth and CO2 emissions nexus in Taiwan using a mixed-frequency VAR model. Economic Analysis and Policy , 79, 319–42.

12.

Dong

, Yang

, Kuang

, He

, Erdal

and Kenski

(2009) PM2.5 concentration prediction using hidden semi-Markov modelbased times series data mining. Expert Systems with Applications , 36, 9046–55.

13.

Economou

and Menary

(2019) A hidden semi-Markov model for characterizing regime shifts in ocean density variability. Journal of the Royal Statistical Society Series C: Applied Statistics , 68, 1529–53.

14.

Economou

, Bailey

and Kapelan

(2014) MCMC implementation for Bayesian hidden semi-Markov models with illustrative applications. Statistics and Computing , 24, 739–52.

15.

Evangelista

, Acal

, Aguilera

, Sarra

, Di Battista

and Palermi

(2024) High dimensional variable selection through group lasso for multiple function-on-function linear regression: A case study in PM10 monitoring. Environmetrics , 36, e2852.

16.

Farcomeni

(2025) Exact score and information matrix for panel hidden semi-Markov models. Statistics and Computing , 35, 55.

17.

Foroni

, Marcellino

and Schumacher

(2015) Unrestricted mixed data sampling (Midas): Midas regressions with unrestricted lag polynomials. Journal of the Royal Statistical Society. Series A (Statistics in Society) , 178, 57–82.

18.

Gåsemyr

and Scheel

(2019) Alternatives to post-processing posterior predictive p values. Scandinavian Journal of Statistics , 46, 1252–73.

19.

Gelman

, Xiao-Li

and Hal

(1996) Posterior predictive assessment of model fitness via realized discrepancies. Statistica Sinica , 6, 733–60.

20.

Ghosh

, Khare

and Michailidis

(2023) The Bayesian nested lasso for mixed frequency regression models. The Annals of Applied Statistics , 17, 2279–2304.

21.

Ghysels

(2016) Macroeconomics and the reality of mixed frequency data. Journal of Econometrics , 193, 294–314.

22.

Ghysels

, Santa-Clara

and Valkanov

(2005) There is a risk-return trade-off after all. Journal of Financial Economics , 76, 509–48. Doi 10.1016/j.jfineco.2004.03.008. URL https://www-sciencedirect-com-443.web.bisu.edu.cn/science/article/pii/S0304405X05000140

23.

Ghysels

, Sinko

and Valkanov

(2007) MIDAS regressions: Further results and new directions. Econometric Reviews , 26, 53–90. doi: 10.1080/07474930600972467.

24.

Guédon

(2003) Estimating hidden semi-Markov chains from discrete sequences. Journal of Computational and Graphical Statistics , 12, 604–39.

25.

Guérin

and Marcellino

(2013) Markovswitching MIDAS models. Journal of Business and Economic Statistics , 31, 45–56. doi: 10.1080/07350015.2012.727721.

26.

Hadj-Amar

, Jewson

and Fiecas

(2023) Bayesian Approximations to Hidden Semi-Markov Models for Telemetric Monitoring of Physical Activity. Bayesian Analysis , 18, 547–77.

27.

Haji-Maghsoudi

, Bulla

, Sadeghifar

, Roshanaei

and Mahjub

(2021) Generalized linear mixed hidden semi-Markov models in longitudinal settings: A Bayesian approach. Statistics in Medicine , 40, 2373–88.

28.

Hanssen-Bauer

(1985) A simple model for diffusion of SO2 in Bergen. Atmospheric Environment (1967) , 19, 415–22.

29.

Ingrassia

and Rocci

(2011) Degeneracy of the EM algorithm for the MLE of multivariate Gaussian mixtures and dynamic constraints. Computational Statistics and Data Analysis , 55, 1715–25.

30.

Jonassen

, Ólafsson

, Valved

, Reuder

and Olseth

(2013) Simulations of the Bergen orographic wind shelter. Tellus A: Dynamic Meteorology and Oceanography , 65, 19206.

31.

Jonghyun

, Tao

and Guanghua

(2014) Bayesian hidden Markov models to identify RNA-protein interaction sites in parclip. Biometrics , 70, 430–40.

32.

Kohns

and Potjagailo

(2025) Flexible Bayesian MIDAS: Time-variation, groupshrinkage and sparsity. Journal of Business and Economic Statistics , pp. 1–17. doi: 10.1080/07350015.2025.2467898.

33.

Lagona

, Maruotti

and Picone

(2011) A non-homogeneous hidden Markov model for the analysis of multi-pollutant exceedance data. In Cappé

, Moulines

and Rydén

(eds) Hidden Markov Models: Theory and Applications , London: Springer, pp. 207–222.

34.

Lin

and Huang

(2025) Penalized composite likelihood estimation for hidden Markov models with unknown number of states. Statistics and Probability Letters , 216, 110247.

35.

Martinez-Zarzoso

and Maruotti

(2013) The environmental Kuznets curve: Functional form, time-varying heterogeneity and outliers in a panel setting. Environmetrics , 24, 461–75.

36.

Maruotti

, Bulla

, Lagona

, Picone

and Martella

(2017) Dynamic mixtures of factor analyzers to characterize multivariate air pollutant exposures. The Annals of Applied Statistics , 11, 1617–48. doi: 10.1214/17-AOAS1049. URL https://doi.org/10.1214/17-AOAS1049

37.

Maruotti

, Punzo

and Bagnato

(2019) Hidden Markov and semi-Markov models with multivariate leptokurtic-normal components for robust modeling of daily returns series. Journal of Financial Econometrics , 17, 91–117.

38.

Maruotti

, Petrella

and Sposito

(2021) Hidden semi-Markov-switching quantile regression for time series. Computational Statistics and Data Analysis , 159, 107–208.

39.

Merlo

, Maruotti

, Petrella

and Punzo

(2022) Quantile hidden semi-Markov models for multivariate time series. Statistics and Computing , 32, 53–90.

40.

Mogliani

and Simoni

(2021) Bayesian MIDAS penalized regressions: Estimation, selection, and prediction. Journal of Econometrics , 222, 833–60.

41.

Pettenuzzo

, Timmermann

and Valkanov

(2016) A MIDAS approach to modeling first and second moment dynamics. Journal of Econometrics , 193, 315–34.

42.

Pohle

, Adam

and Beumer

(2022) Flexible estimation of the state dwell-time distribution in hidden semi-Markov models. Computational Statistics and Data Analysis , 172, 107479.

43.

Robert

, Ryden

and Titterington

(2000) Bayesian inference in hidden Markov models through the reversible jump Markov Chain Monte Carlo method. Journal of the Royal Statistical Society (Series B) , 62, 57–75.

44.

Rodriguez

and Puggioni

(2010) Mixed frequency models: Bayesian approaches to estimation and prediction. International Journal of Forecasting , 26, 293–311.

45.

Russo

and Farcomeni

(2024) A copula formulation for multivariate latent Markov models. TEST , 33, 731–51.

46.

Russo

, Farcomeni

, Pittau

and Zelli

(2024) Covariate-modulated rectangular latent Markov models with an unknown number of regime profiles. Statistical Modeling , 24, 368–88.

47.

Scott

(2002) Bayesian methods for hidden Markov models. Journal of the American Statistical Association , 97, 337–51.

48.

Shi

, Punzo

, Otneim

and Maruotti

(2025) Hidden semi-Markov models for rainfall-related insurance claims. Insurance: Mathematics and Economics , 120, 91–106. doi: 10.1016/j.insmatheco.2024.11.008. URL https://www-sciencedirect-com-443.web.bisu.edu.cn/science/article/pii/S0167668724001136

49.

Spezia

(2010) Bayesian analysis of multivariate Gaussian hidden Markov models with an unknown number of regimes. Journal of Time Series Analysis , 31, 1–11.

50.

Watanabe

(2010) Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory. Journal of Machine Learning Research , 11, 3571–94.

51.

Wolf

, Esau

and Reuder

(2014) Analysis of the vertical temperature structure in the Bergen valley, Norway, and its connection to pollution episodes. Journal of Geophysical Research: Atmospheres , 119, 10–645.

52.

Wolf-Grosse

, Esau

and Reuder

(2017) The large-scale circulation during air quality hazards in Bergen, Norway. Tellus A: Dynamic Meteorology and Oceanography , 69, 1406265.

53.

S-Z

(2010) Hidden semi-Markov models. Artificial Intelligence , 174, 215–243.

54.

Zhao

, Han

, Ding

and Calin

(2018) Forecasting carbon dioxide emissions based on a hybrid of mixed data sampling regression model and back propagation neural network in the USA. Environmental Science and Pollution Research , 25, 2899–2910.

55.

Zucchini

and MacDonald

(2009) Hidden Markov Models for Time Series: An Introduction Using R. London: Chapman and Hall/CRC.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.34 MB

0.00 MB