pith. sign in

arxiv: 2603.27142 · v2 · submitted 2026-03-28 · 📊 stat.ML · cs.AI· cs.LG

tBayes-MICE: A Bayesian Approach to Multiple Imputation for Time Series Data

Amuche Ibenegbu , Pierre Lafaye de Micheaux , Rohitash Chandra This is my paper

Pith reviewed 2026-05-14 22:25 UTC · model grok-4.3

classification 📊 stat.ML cs.AIcs.LG
keywords time-series imputationBayesian MICEmultiple imputationMCMC samplingmissing datatemporal dependenciesair qualityphysiological monitoring
0
0 comments X

The pith

tBayes-MICE extends MICE with Bayesian MCMC sampling and time-lagged features to reduce imputation errors in time-series data while quantifying uncertainty.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops tBayes-MICE as a Bayesian extension of Multiple Imputation by Chained Equations for handling missing values in time-series data. By adding time-lagged features and using Markov Chain Monte Carlo sampling, the method accounts for uncertainty in both the imputation models and the filled-in values. Tests on real datasets from air quality monitoring and physiological signals demonstrate lower imputation errors than standard approaches across all variables examined. The work also compares two MCMC samplers and finds that the Metropolis-Adjusted Langevin Algorithm mixes more effectively than random walk Metropolis in most cases.

Core claim

The central discovery is that tBayes-MICE, by embedding Bayesian inference and MCMC sampling into the MICE procedure along with temporal lags, achieves lower imputation errors than baseline methods on the AirQuality and PhysioNet datasets while properly incorporating uncertainty.

What carries the argument

Bayesian MICE with MCMC sampling and time-lagged features, which performs posterior sampling over conditional imputation models to handle temporal dependencies in missing data.

If this is right

  • Imputation errors are reduced relative to baseline methods for every variable in the evaluated datasets.
  • Uncertainty in the imputation process is quantified through posterior distributions, providing a more accurate assessment of error.
  • The MALA sampler shows superior mixing compared to RWM across most variables with similar accuracy.
  • The approach offers a practical method for time-series imputation in environmental and clinical applications.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This could improve the reliability of time-series predictions in fields where missing data is common by feeding the imputed data with uncertainty into forecasting models.
  • Future work might test the method on datasets with different missing patterns or higher dimensions.
  • The Bayesian framework could be combined with other imputation techniques like deep learning for even better performance.

Load-bearing premise

The time-lagged features and MCMC sampling on the conditional models will accurately reflect the temporal dependencies without introducing bias or failing to converge properly.

What would settle it

A direct comparison showing that tBayes-MICE imputation errors exceed those of standard MICE on the same AirQuality and PhysioNet test data, or that the uncertainty estimates do not align with actual imputation inaccuracies.

Figures

Figures reproduced from arXiv: 2603.27142 by Amuche Ibenegbu, Pierre Lafaye de Micheaux, Rohitash Chandra.

Figure 1
Figure 1. Figure 1: Time-lagged imputation mechanism in univariate MICE, where a missing value at time index t is imputed by conditioning on observed neigh￾bouring values from past and future lags. Assuming that the data are MAR, we present the time-lagged FCS procedure in Algorithm 1. Algorithm 1 Time Series MICE 1: Input: Incomplete original dataset X 0 with p time-indexed variables, lag orders (ℓp, ℓ f), from which we deri… view at source ↗
Figure 2
Figure 2. Figure 2: The proposed tBayes-MICE imputation framework. The framework consists of temporal pattern detection, placeholder initialisation, MICE loop with lagged predictors and Bayesian modelling, while MCMC (RWM or MALA) provides posterior sampling and parameter updates. The final imputation is generated via posterior predictive draws. plers have nearly identical posterior means and uncertainty in￾tervals for all of… view at source ↗
Figure 3
Figure 3. Figure 3: Convergence diagnostics for τ 2 (trace, marginal density, and ACF) across two chains and two samplers for one of the variables (HC03) from the physionet dataset [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Prediction accuracy comparison for the CO(GT) variable from the AirQuality dataset (top) and the HCO3 variable from the physioNet dataset (bottom). Each subplot panel shows:(top-left) predicted versus true values, (top-right) residuals against the true values, (bottom-left) sorted true values with corresponding model predictions, and (bottom-right) box plots of absolute errors. 12 [PITH_FULL_IMAGE:figures… view at source ↗
Figure 5
Figure 5. Figure 5: Imputation error patterns over time index of each method across different datasets, with the AirQuality CO(GT) variable shown in the top panel and the PhysioNet HCO3 variable shown in the bottom panel 14 [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Posterior predictive performance of the proposed tBayes-MICE model over time for the environmental CO(GT) variable (top row) and the clinical Na variable (bottom row), shown across training(left) and test(right) sets. The blue lines represent observed values, the red lines indicate posterior predictive means, and the shaded regions denote the 95% credible intervals [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Comparative imputation accuracy and distribution across runs (box plot) for deterministic and Bayesian methods on the PhsyioNet dataset using NRMSE provides additional stability for temporally structured variables and across datasets. Accordingly, we adopt tBayes-MICE-V2 as the default configuration in this study. 5. Discussion Our study introduces tBayes-MICE, a Bayesian enhancement of the Multiple Imputa… view at source ↗
read the original abstract

Time-series analysis is often affected by missing data, a common problem across several fields, including healthcare and environmental monitoring. Multiple Imputation by Chained Equations (MICE) has been prominent for imputing missing values through "fully conditional specification". We extend MICE using the Bayesian framework (tBayes-MICE), utilising Bayesian inference to impute missing values via Markov Chain Monte Carlo (MCMC) sampling to account for uncertainty in MICE model parameters and imputed values. We also include temporally informed initialisation and time-lagged features in the model to respect the sequential nature of time-series data. We evaluate the tBayes-MICE method using two real-world datasets (AirQuality and PhysioNet), and using both the Random Walk Metropolis (RWM) and the Metropolis-Adjusted Langevin Algorithm (MALA) samplers. Our results demonstrate that tBayes-MICE reduces imputation errors relative to the baseline methods over all variables and accounts for uncertainty in the imputation process, thereby providing a more accurate measure of imputation error. We also found that MALA mixed better than RWM across most variables, achieving comparable accuracy while providing more consistent posterior exploration. Overall, these findings suggest that the tBayes-MICE framework represents a practical and efficient approach to time-series imputation, balancing increased accuracy with meaningful quantification of uncertainty in various environmental and clinical settings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper proposes tBayes-MICE, a Bayesian extension of Multiple Imputation by Chained Equations (MICE) for time-series data that uses MCMC sampling (RWM and MALA) on conditional models augmented with time-lagged features and temporally informed initialization to account for parameter and imputation uncertainty. It evaluates the method on the AirQuality and PhysioNet datasets and claims reduced imputation errors relative to baselines across all variables together with improved uncertainty quantification.

Significance. If the MCMC sampling converges reliably and the empirical reductions hold, the framework could provide a practical Bayesian alternative to standard MICE for sequential data in healthcare and environmental monitoring, offering both point imputations and posterior-based uncertainty estimates that are currently missing from many chained-equation approaches.

major comments (3)
  1. [Abstract] Abstract and evaluation sections: the central claim that tBayes-MICE 'reduces imputation errors relative to the baseline methods over all variables' is stated without any reported numerical values, error bars, baseline definitions, or per-variable metrics, so the support for the primary result cannot be assessed from the provided evidence.
  2. [MCMC Sampling] MCMC implementation and results: no R-hat statistics, effective sample sizes, trace plots, or autocorrelation times are supplied for the conditional posteriors on either dataset, yet the uncertainty quantification and error-reduction claims rest entirely on the MCMC samples faithfully representing the target posterior.
  3. [Method] Method description: the assumption that time-lagged features plus MCMC on the fully conditional specification will capture temporal dependencies without bias or convergence failure is not accompanied by any diagnostic checks or sensitivity analysis, leaving open the possibility that reported improvements are artifacts of poor mixing.
minor comments (2)
  1. [Abstract] The abstract refers to 'over all variables' without listing the variables or providing variable-wise breakdowns, which reduces clarity.
  2. [Method] Notation for the Bayesian conditional models and the precise form of the time-lagged feature augmentation could be made more explicit to aid reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their thorough and constructive review. We have addressed each major comment below and will revise the manuscript to strengthen the presentation of results, add necessary diagnostics, and provide additional analyses as outlined.

read point-by-point responses

  1. Referee: [Abstract] Abstract and evaluation sections: the central claim that tBayes-MICE 'reduces imputation errors relative to the baseline methods over all variables' is stated without any reported numerical values, error bars, baseline definitions, or per-variable metrics, so the support for the primary result cannot be assessed from the provided evidence.

    Authors: We agree that the abstract and evaluation sections would be strengthened by including specific quantitative results. In the revised manuscript, we will report numerical imputation error values (e.g., mean RMSE and MAE) for tBayes-MICE versus the baselines, including standard deviations to serve as error bars. We will explicitly define the baseline methods and add a table or figure with per-variable metrics for both datasets to allow readers to fully assess the primary claims. revision: yes

  2. Referee: [MCMC Sampling] MCMC implementation and results: no R-hat statistics, effective sample sizes, trace plots, or autocorrelation times are supplied for the conditional posteriors on either dataset, yet the uncertainty quantification and error-reduction claims rest entirely on the MCMC samples faithfully representing the target posterior.

    Authors: We acknowledge this gap in the reporting of MCMC diagnostics. We will add R-hat statistics, effective sample sizes, and autocorrelation times for the conditional posteriors on both the AirQuality and PhysioNet datasets. Representative trace plots and summary convergence diagnostics will be included in the main text or supplementary material to confirm that the samples faithfully represent the target posteriors for both RWM and MALA samplers. revision: yes

  3. Referee: [Method] Method description: the assumption that time-lagged features plus MCMC on the fully conditional specification will capture temporal dependencies without bias or convergence failure is not accompanied by any diagnostic checks or sensitivity analysis, leaving open the possibility that reported improvements are artifacts of poor mixing.

    Authors: We will revise the method section to include explicit diagnostic checks and sensitivity analyses. This will encompass the convergence metrics noted above, plus sensitivity tests on the number of time lags and initialization strategies. These additions will demonstrate that temporal dependencies are captured reliably and that improvements are not artifacts of poor mixing or convergence issues. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The paper extends standard MICE with Bayesian MCMC sampling (RWM/MALA) and time-lagged features for time-series imputation. No equations, predictions, or central claims reduce by construction to fitted inputs or self-citations; the method is presented as a direct application of established fully conditional specification and MCMC techniques on external datasets (AirQuality, PhysioNet). No self-definitional loops, fitted-input predictions, or load-bearing self-citations appear in the derivation. The approach remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no explicit free parameters, axioms, or invented entities; the approach rests on standard Bayesian priors and MCMC assumptions whose details are not stated.

pith-pipeline@v0.9.0 · 5555 in / 970 out tokens · 34219 ms · 2026-05-14T22:25:22.538264+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

60 extracted references · 60 canonical work pages · 1 internal anchor

  1. [1]

    Afrifa-Yamoah, U

    E. Afrifa-Yamoah, U. A. Mueller, S. M. Taylor, and A.J. Fisher. Missing data imputation of high-resolution tempo- ral climate time series data.Meteorological Applications, 27(1):e1873, 2020

    work page 2020

  2. [2]

    M.J. Azur, E. A. Stuart, C. Frangakis, and P. J. Leaf. Mul- tiple imputation by chained equations: what is it and how does it work?International Journal of Methods in Psy- chiatric Research, 20(1):40–49, 2011

    work page 2011

  3. [3]

    G. E. Batista, M. C. Monard, et al. A study of k-nearest neighbour as an imputation method.His, 87(251-260):48, 2002

    work page 2002

  4. [4]

    Brand.Development, implementation and evaluation of multiple imputation strategies for the statistical analysis of incomplete data sets

    J. Brand.Development, implementation and evaluation of multiple imputation strategies for the statistical analysis of incomplete data sets. Thesis, 1999

    work page 1999

  5. [5]

    W. Cao, D. Wang, J. Li, H. Zhou, L. Li, and Y . Li. Brits: Bidirectional recurrent imputation for time series. Advances in Neural Information Processing Systems, 31, 2018

    work page 2018

  6. [6]

    Chandra and J

    R. Chandra and J. Simmons. Bayesian neural networks via mcmc: a python-based tutorial.IEEE Access, 12:70519– 70549, 2024

    work page 2024

  7. [7]

    Z. Che, S. Purushotham, K. Cho, D. Sontag, and Y . Liu. Recurrent neural networks for multivariate time series with missing values.Scientific Reports, 8(1):6085, 2018

    work page 2018

  8. [8]

    De Vito, E

    S. De Vito, E. Massera, M. Piga, L. Martinotto, and G. Di Francia. On field calibration of an electronic nose for benzene estimation in an urban pollution mon- itoring scenario.Sensors and Actuators B: Chemical, 129(2):750–757, 2008

    work page 2008

  9. [9]

    A R. T. Donders, G.JMG. Van Der Heijden, T. Stijnen, and K. Moons. A gentle introduction to imputation of missing values.Journal of Clinical Epidemiology, 59(10):1087– 1091, 2006

    work page 2006

  10. [10]

    Dudhe, M

    P. Dudhe, M. H. Li, J.and Yousuf, J. Lehmann, and S. Vah- dati. Icu mechanical ventilation data imputation. InIFIP International Conference on Artificial Intelligence Appli- cations and Innovations, pages 140–154. Springer, 2025

    work page 2025

  11. [11]

    N. S. Erler, D. Rizopoulos, V . W.V . Jaddoe, O. H. Franco, and E. MEH Lesaffre. Bayesian imputation of time- varying covariates in linear mixed models.Statistical Methods in Medical Research, 28(2):555–568, 2019

    work page 2019

  12. [12]

    S. Fang, Q. Wen, Y . Luo, S. Zhe, and L. Sun. Bay- otide: Bayesian online multivariate time series impu- tation with functional decomposition.arXiv preprint arXiv:2308.14906, 2023

    work page arXiv 2023

  13. [13]

    Farhangfar, L

    A. Farhangfar, L. A. Kurgan, and W. Pedrycz. A novel framework for imputation of missing values in databases. IEEE Transactions on Systems, Man, and Cybernetics- Part A: Systems and Humans, 37(5):692–709, 2007

    work page 2007

  14. [14]

    P. J. García-Laencina and A. R. Sancho-Gómez, J. L.and Figueiras-Vidal. Pattern classification with missing data: a review.Neural Computing and Applications, 19:263– 282, 2010

    work page 2010

  15. [15]

    A. E. Gelfand and A. FM. Smith. Sampling-based ap- proaches to calculating marginal densities.Journal of the American Statistical Association, 85(410):398–409, 1990

    work page 1990

  16. [16]

    Gelman, W

    A. Gelman, W. R. Gilks, and G. O. Roberts. Weak con- vergence and optimal scaling of random walk metropolis algorithms.The Annals of Applied Probability, 7(1):110– 120, 1997

    work page 1997

  17. [17]

    Gelman, G

    A. Gelman, G. O. Roberts, and W. R. Gilks. Efficient metropolis jumping rules.Bayesian statistics 5, 5:599– 608, 1996

    work page 1996

  18. [18]

    Gelman and D

    A. Gelman and D. B. Rubin. Inference from iterative simulation using multiple sequences.Statistical Science, 7(4):457–472, 1992

    work page 1992

  19. [19]

    Geman and D

    S. Geman and D. Geman. Stochastic relaxation, gibbs dis- tributions, and the bayesian restoration of images.IEEE Transactions on Pattern Analysis and Machine Intelli- gence, (6):721–741, 1984

    work page 1984

  20. [20]

    W. R. Gilks, S. Richardson, and D. Spiegelhalter.Markov chain Monte Carlo in practice. CRC press, 1995

    work page 1995

  21. [21]

    Grzesiak, C

    K. Grzesiak, C. Muller, J. Josse, and J. Näf. Do we need dozens of methods for real world missing value imputa- tion?arXiv preprint arXiv:2511.04833, 2025

    work page arXiv 2025

  22. [22]

    Haario, E

    H. Haario, E. Saksman, and J. Tamminen. An adaptive metropolis algorithm.Bernoulli Society for Mathematical Statistics and Probability, 7(2):223–159, 2001

    work page 2001

  23. [23]

    W. K. Hastings. Monte carlo sampling methods using markov chains and their applications. 1970. 18

    work page 1970

  24. [24]

    V . Hua, T. Nguyen, M. Dao, H. Nguyen, and B. T. Nguyen. The impact of data imputation on air quality pre- diction problem.Plos one, 19(9):e0306303, 2024

    work page 2024

  25. [25]

    Jäger, A

    S. Jäger, A. Allhorn, and F. Bießmann. A benchmark for data imputation methods.Frontiers in Big Data, 4:693674, 2021

    work page 2021

  26. [26]

    L. Ji, M. Chen, Z. Oravecz, E. M. Cummings, Z. Lu, and S. Chow. A bayesian vector autoregressive model with nonignorable missingness in dependent variables and covariates: Development, evaluation, and application to family processes.Structural Equation Modeling: A Mul- tidisciplinary Journal, 27(3):442–467, 2020

    work page 2020

  27. [27]

    Junninen, H

    H. Junninen, H. Niska, K. Tuppurainen, J. Ruuskanen, and M. Kolehmainen. Methods for imputation of missing values in air quality data sets.Atmospheric Environment, 38(18):2895–2907, 2004

    work page 2004

  28. [28]

    S. I. Khan and A. S. Md L. Hoque. Sice: an improved missing data imputation technique.Journal of Big Data, 7(1):37, 2020

    work page 2020

  29. [29]

    Kulkarni and R

    O. Kulkarni and R. Chandra. Bayes-catsi: A variational bayesian deep learning framework for medical time series data imputation.arXiv preprint arXiv:2410.01847, 2024

    work page arXiv 2024

  30. [30]

    Le Morvan, J

    M. Le Morvan, J. Josse, E. Scornet, and G. Varoquaux. What’sa good imputation to predict with missing val- ues?Advances in Neural Information Processing Systems, 34:11530–11540, 2021

    work page 2021

  31. [31]

    John Wiley & Sons, 2019

    R.JA Little and D.B Rubin.Statistical analysis with miss- ing data, volume 793. John Wiley & Sons, 2019

    work page 2019

  32. [32]

    Metropolis, A

    N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller, and E. Teller. Equation of state calculations by fast computing machines.The Journal of Chemical Physics, 21(6):1087–1092, 1953

    work page 1953

  33. [33]

    Comparison of different Methods for Univariate Time Series Imputation in R

    S. Moritz, T. Sardá, A.and Bartz-Beielstein, M. Zaef- ferer, and J. Stork. Comparison of different methods for univariate time series imputation in r.arXiv preprint arXiv:1510.03924, 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  34. [34]

    J. S. Murray. Multiple imputation: a review of practical and theoretical findings.Statistical Science, 33:142–159, 2018

    work page 2018

  35. [35]

    J. Näf, E. Scornet, and J. Josse. What is a good imputation under mar missingness?arXiv preprint arXiv:2403.19196, 2024

    work page arXiv 2024

  36. [36]

    N. M. Noor, M. M. Al Bakri Abdullah, A. S. Yahaya, and N. A. Ramli. Comparison of linear interpolation method and mean method to replace the missing values in envi- ronmental data set. InMaterials Science Forum, volume 803, pages 278–281. Trans Tech Publ, 2015

    work page 2015

  37. [37]

    Resche-Rigon and I

    M. Resche-Rigon and I. R. White. Multiple imputation by chained equations for systematically and sporadically missing multilevel data.Statistical Methods in Medical Research, 27(6):1634–1649, 2018

    work page 2018

  38. [38]

    C. P. Robert, G. Casella, and G. Casella.Monte Carlo statistical methods, volume 2. Springer, 1999

    work page 1999

  39. [39]

    G. O. Roberts and J. S. Rosenthal. Optimal scaling of dis- crete approximations to langevin diffusions.Journal of the Royal Statistical Society: Series B (Statistical Methodol- ogy), 60(1):255–268, 1998

    work page 1998

  40. [40]

    G. O. Roberts and J. S. Rosenthal. Optimal scaling for var- ious metropolis-hastings algorithms.Statistical Science, 16(4):351–367, 2001

    work page 2001

  41. [41]

    G. O. Roberts and R. L. Tweedie. Exponential conver- gence of langevin distributions and their discrete approxi- mations. 1996

    work page 1996

  42. [42]

    D. B. Rubin. An overview of multiple imputation. In Proceedings of the survey research methods section of the American Statistical Association, volume 79, page 84, 1988

    work page 1988

  43. [43]

    D.B. Rubin. Multiple imputations in sample surveys-a phenomenological bayesian approach to nonresponse. In Proceedings of the survey research methods section of the American Statistical Association, volume 1, pages 20–34. American Statistical Association Alexandria, V A, 1978

    work page 1978

  44. [44]

    C. M. Salgado, C. Azevedo, H. Proença, and S. M. Vieira. Missing data.Secondary Analysis of Electronic Health Records, pages 143–162, 2016

    work page 2016

  45. [45]

    J. L. Schafer.Analysis of incomplete multivariate data. CRC press, 1997

    work page 1997

  46. [46]

    J.L. Schafer. Multiple imputation: a primer.Statistical Methods in Medical Research, 8(1):3–15, 1999

    work page 1999

  47. [47]

    Shadbahr, M

    T. Shadbahr, M. Roberts, J. Stanczuk, J. Gilbey, P. Teare, S. Dittmer, M. Thorpe, R. V . Torné, E. Sala, P. Lió, et al. The impact of imputation quality on machine learning classifiers for datasets with missing values.Communica- tions Medicine, 3(1):139, 2023

    work page 2023

  48. [48]

    Silva, G

    I. Silva, G. Moody, D. J. Scott, L. A. Celi, and R. G. Mark. Predicting in-hospital mortality of icu patients: The phy- sionet/computing in cardiology challenge 2012. In2012 Computing in Cardiology, pages 245–248. IEEE, 2012

    work page 2012

  49. [49]

    D. J. Stekhoven and P. Bühlmann. Missforest—non- parametric missing value imputation for mixed-type data. Bioinformatics, 28(1):112–118, 2012

    work page 2012

  50. [50]

    W. Sun. Application of markov chain monte-carlo mul- tiple imputation method to deal with missing data from the mechanism of mnar in sensitivity analysis for a lon- gitudinal clinical trial. InMonte-Carlo Simulation-Based Statistical Modeling, pages 233–252. Springer, 2017. 19

    work page 2017

  51. [51]

    M. A. Tanner and W. H. Wong. The calculation of pos- terior distributions by data augmentation.Journal of the American statistical Association, 82(398):528–540, 1987

    work page 1987

  52. [52]

    Van Buuren

    S. Van Buuren. Flexible multivariate imputation by mice. TNO Prevention and Health, 1999

    work page 1999

  53. [53]

    van Buuren.Flexible Imputation of Missing Data

    S. van Buuren.Flexible Imputation of Missing Data. CRC Press, 2018

    work page 2018

  54. [54]

    Van Buuren, H

    S. Van Buuren, H. C. Boshuizen, and D. L. Knook. Mul- tiple imputation of missing blood pressure covariates in survival analysis.Statistics in Medicine, 18(6):681–694, 1999

    work page 1999

  55. [55]

    Van Buuren and K

    S. Van Buuren and K. Groothuis-Oudshoorn. mice: Mul- tivariate imputation by chained equations in r.Journal of Statistical Software, 45:1–67, 2011

    work page 2011

  56. [56]

    W. F. Velicer and S. M. Colby. A comparison of missing- data procedures for arima time-series analysis.Educa- tional and Psychological Measurement, 65(4):596–615, 2005

    work page 2005

  57. [57]

    I. R. White and A.M. Royston, P.and Wood. Multiple im- putation using chained equations: issues and guidance for practice.Statistics in Medicine, 30(4):377–399, 2011

    work page 2011

  58. [58]

    K. Yin, L. Feng, and W. K. Cheung. Context-aware time series imputation for multi-analyte clinical data.Journal of Healthcare Informatics Research, 4(4):411–426, 2020

    work page 2020

  59. [59]

    Gain: Missing data imputation using generative adversarial nets

    Jinsung Yoon, James Jordon, and Mihaela Schaar. Gain: Missing data imputation using generative adversarial nets. InInternational Conference on Machine Learning, pages 5689–5698. PMLR, 2018

    work page 2018

  60. [60]

    Yozgatligil, S

    C. Yozgatligil, S. Aslan, C. Iyigun, and I. Batmaz. Com- parison of missing value imputation methods in time se- ries: the case of turkish meteorological data.Theoretical and Applied Climatology, 112(1):143–167, 2013. 20

    work page 2013