Pith. sign in

REVIEW 3 major objections 5 minor 31 references

Enhancing Bagging Ensemble Regression with Data Integration for Time Series-Based Diabetes Prediction

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An enhanced bagging ensemble with stratified block bootstrap and inverse-RMSE weighting beats six baselines for U.S. state-level diabetes prevalence in 2021

desk verdict Useful integrated dataset and a plausible bagging tweak, but the headline accuracy gains rest on test-set lag selection and need error bars before they hold. read the letter →

arxiv 2506.13786 v1 pith:DAX4AM3O submitted 2025-06-11 cs.LG cs.AI

Vuong M. Ngo , Tran Quang Vinh , Patricia Kearney , Mark Roantree This is my paper · ORCID
classification cs.LGcs.AI
keywords diabetesprevalencebaggingensembleregressiontimeseriesforecastingstratifiedblockbootstrapinverse-RMSEweightinghealthdataintegrationstate-levelpredictionmachinelearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that a modified bagging ensemble regression model, called ERMBag+ in the main text, predicts next-year diabetes prevalence for U.S. states more accurately than six standard machine learning baselines. It builds a 90-feature dataset by integrating four U.S. public data sources spanning 2011 to 2021, then converts the panel into supervised time-series examples using lagged predictors. The model's two main changes to ordinary bagging are stratified block bootstrap sampling, which preserves temporal ordering and subgroup distributions, and weighted aggregation in which each tree's vote counts inversely to its validation RMSE. On the held-out year 2021 across 51 states, the paper reports ERMBag+ achieves MAE = 0.41, RMSE = 0.53, MAPE = 4.01, and $R^2 = 0.91$, the best among all models compared. If correct, this gives public-health planners a practical recipe for state-level diabetes forecasting from routinely collected public data.

What carries the argument

The load-bearing mechanism is the ERMBag+ training procedure: stratified block bootstrap sampling for time-series data, early stopping per decision tree, and weighted aggregation with $w_i = 1/\mathrm{RMSE}(f_i)$ normalized over the ensemble. Lag-window construction turns the panel of 89 annual features into supervised examples with $89 \times (l+1) + l$ variables, so that at the selected Lag=2 each example covers the current and two preceding years. The block bootstrap is supposed to preserve temporal autocorrelation while retaining demographic and economic subgroup proportions, and the inverse-RMSE weights let better-validated trees dominate the final forecast.

What would settle it

Replace the interpolated 2021 population and housing values with the actual 2021 Census counts and rerun the full comparison with Lag chosen by cross-validation on 2011 to 2020 data; if the reported MAE = 0.41 and $R^2 = 0.91$ persist, the central claim is supported, while if the gap over ERMBag narrows sharply, the reported advantage depends on interpolation and test-set tuning.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that replacing plain bootstrap resampling with stratified block bootstrap and replacing equal averaging with inverse-RMSE weighted averaging turns an ordinary bagged decision-tree regressor into the most accurate model for the 2021 diabetes-prevalence forecasts. The reported winner is ERMBag+ at Lag=2, where each training example concatenates 89 predictors from the current and two previous years plus the two previous target values, yielding $89 \times (l+1) + l = 269$ variables per example. Its reported error metrics are MAE = 0.41, RMSE = 0.53, MAPE = 4.01, and $R^2 = 0.91$, compared with the runner-up ERMBag at MAE = 0.44, RMSE = 0.60, MAPE = 4.38, and $R^2 = 0.88$. The paper attributes the gain to better preservation of temporal dependencies and to giving more weight to trees that validate well.

Load-bearing premise

The load-bearing premise is that the 2021 population and housing features are accurate enough despite being linearly interpolated rather than observed, and that selecting each model's best Lag from the 2021 test results does not inflate the reported advantage.

Editorial extensions

If this is right

  • ERMBag+ at Lag=2 yields a one-year-ahead state-level forecast with reported MAE of 0.41 percentage points and $R^2 = 0.91$, the best accuracy among all models the paper compares.
  • Because ERMBag+ differs from the ERMBag baseline by stratified block bootstrap, early stopping, and inverse-RMSE weighting, the reported gap isolates the benefit of those modifications.
  • The lag analysis shows that the best window length is model-dependent, with Lag=1 for BDTree and LSBoost, Lag=2 for the bagging variants, Lag=6 for SVMReg, and Lag=9 for LSTM, so practitioners should treat lag as a hyperparameter.
  • The integrated 90-feature dataset is assembled from public annual sources, so the same preprocessing pipeline can be rerun as CDC, Census, BLS, and BEA data are updated.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper reports each model's best Lag chosen from the test-year results; a stricter protocol would tune Lag inside the training period only, and rerunning the comparison that way could shrink but would not necessarily erase the reported gap.
  • The 71 chronic-disease indicators likely carry much of the predictive signal, since they include direct health-behavior and comorbidity measures; an ablation by feature category would test whether demographic, housing, and economic features add accuracy beyond those indicators.
  • An $R^2$ of 0.91 with MAPE of 4.01 implies average errors near 0.4 percentage points, which can still be material in states with low diabetes prevalence; reporting per-state error bounds would sharpen the policy use of the forecasts.
  • The same stratified block bootstrap plus inverse-RMSE weighting recipe could be transferred to other annual state-level health outcomes, such as obesity or kidney disease, using the same public data pipeline.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper integrates four U.S. government datasets (chronic disease indicators, population, housing, and economy) into a 90-feature state-year panel from 2011 to 2021, and proposes ERMBag+, a bagging ensemble of decision trees with stratified block bootstrap, early stopping, and inverse-RMSE weighted aggregation. It evaluates ERMBag+ against six baselines (SVMReg, BDTree, LSBoost, NN, LSTM, ERMBag) for one-step-ahead prediction of state-level diabetes prevalence in 2021, and reports that ERMBag+ achieves the best metrics (MAE = 0.41, RMSE = 0.53, MAPE = 4.01, R² = 0.91).

Significance. If the empirical comparison were valid, the paper would make a useful applied contribution by demonstrating a reproducible data-engineering pipeline for state-level health forecasting and a modestly improved bagging variant. The data integration across four public sources and the explicit time-series framing are constructive. However, the central claim of superiority is currently supported only by a single test year with 51 observations, a test-set-based selection of lag hyperparameters, and no uncertainty quantification; the methodological novelty of ERMBag+ over standard bagging is incremental, and the evaluation protocol needs substantial rework before the headline result can be accepted.

major comments (3)
  1. [Section 5.2 and Table 2] The reported performance of every model is the best result over nine lag configurations, with the lag chosen on the same 51-point 2021 test set used for the final comparison. This is test-set selection: the headline ERMBag+ figures (MAE = 0.41, RMSE = 0.53, R² = 0.91) are the maximum over Lag = 1,...,9 of in-sample test performance, not an unbiased out-of-sample estimate. Because the margins over ERMBag are small (ΔMAE = 0.03, ΔRMSE = 0.07), and 51 points cannot reliably discriminate among nine correlated configurations, this protocol could plausibly manufacture the reported ranking. The authors should select Lag and any other hyperparameters by temporal (rolling-origin) validation on 2011-2020, then evaluate once on 2021, or report results for all lags without per-model selection.
  2. [Section 5.2 and Algorithm 1] The test set contains only 51 observations, and no confidence intervals, standard deviations across repeated runs, or significance tests are provided. NN, LSTM, and the bagging ensembles are stochastic, so single-run comparisons are not sufficient to support the claim that ERMBag+ outperforms ERMBag and NN. Additionally, several hyperparameters of ERMBag+ are not specified: the number of base learners M, the block size B, the early-stopping threshold, and the validation split used to compute RMSE weights in Step 3 and Step 4 of Algorithm 1. Without these details, the contribution of each proposed enhancement cannot be isolated, and the experiments are not reproducible.
  3. [Section 3.1] The 2021 values for population and housing features are not observed but are obtained by linear interpolation (population from 2011-2020 trends and the 2021 Census; housing from 2010 and 2020 decennial data). Since 2021 is exactly the test year, every model prediction rests on interpolated inputs; if the interpolation error is correlated with diabetes prevalence or with state-level trends, the evaluation could be systematically biased. The paper should either validate the interpolation against held-out observed years, or repeat the experiment using an earlier year with fully observed inputs as the test period.
minor comments (5)
  1. [Abstract and Section 5.2] The abstract reports an R² of 0.9 while Section 5.2 reports R² = 0.91; the inconsistent number should be reconciled.
  2. [Abstract and Section 4.4] The model name is written as 'EBMBag+' in most of the paper but 'ERMBag+' in the abstract and in Table 2; the naming should be made consistent.
  3. [Section 4.4, Algorithm 1] The 'stratified block bootstrap' is not fully defined: it is not specified which variables are used for stratification, how the non-overlapping time blocks are formed when lagged features are present, or how the block size B is chosen. A precise algorithmic description is needed.
  4. [Section 5.2, Figure 3] Figure 3 is described as showing average prediction performance, but no error bars or dispersion measures are provided; for stochastic models this makes the visual comparisons uninformative.
  5. [Section 5.2] The statement that ERMBag+ outperforms baselines by '3.4% to 61.1%' is not backed by an explicit calculation in the paper and should be derived from the metric formulas or replaced by directly reported relative differences.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: ERMBag+ is a substantive variant and the comparison is computed on held-out 2021 data; test-set lag selection and underspecified validation are correctness risks, not circular steps.

full rationale

The paper's derivation chain is not circular. Training inputs are the 89 predictor features and lagged targets from 2011-2020, while the 2021 diabetes prevalence is used only as the test label after model construction; no target value is baked into the feature construction or the model definition. ERMBag+ is a substantive algorithmic variant (stratified block bootstrap, early stopping on validation error, inverse-RMSE weighted averaging), not a renamed existing predictor, and the baseline comparisons are computed on held-out 2021 state-year observations. The only author self-citations (refs 22 and 23 in Sections 3.2 and 4.1) justify the data-integration pipeline, which is not load-bearing for the accuracy claim; they are not used to prove uniqueness or to forbid alternative models. Two evaluation weaknesses exist but are not circularity: the Lag in Table 2 is chosen as each model's best value on the 2021 test set, and the inverse-RMSE weights are described as coming from 'validation datasets' whose split is not specified. These are test-set-selection and reproducibility risks, not cases where the prediction reduces to its inputs by construction. Accordingly, no circular step can be exhibited from the text, and the derivation is self-contained for its stated modeling claim.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests less on invented entities than on unspecified hyperparameters and on the assumption that interpolated 2021 values and lagged features capture the forecasting signal. The model itself is a composition of known techniques.

free parameters (4)
  • Lag value per model = ERMBag+ uses 2; baselines use 1 to 9
    Table 2 reports each model's best result over Lag 1 to 9, so the lag is selected on the test set rather than fixed a priori.
  • Number of base learners M = Not reported
    Algorithm 1 requires M, but no value or tuning procedure is given.
  • Block size B = Not reported
    Algorithm 1 requires block size for stratified block bootstrap; the paper only gives qualitative guidance about adjusting it.
  • Early stopping threshold = Not reported
    Step 3 of Algorithm 1 uses an early stopping criterion based on RMSE, but the criterion and the validation split are unspecified.
assumptions (4)
  • domain assumption The supervised lag-based framing can capture the temporal dynamics of diabetes prevalence.
    Section 4.2 assumes that lags 1 to 9 of the 89 features and past prevalence are sufficient; no other temporal structure is modeled.
  • domain assumption Interpolated 2021 population and housing figures are accurate enough for forecasting.
    Section 3.1 states 2021 population and housing data were linearly interpolated, and these estimates feed the test-year inputs.
  • domain assumption CDC CDI diabetes prevalence values are reliable ground truth for each state.
    Test labels for 2021 come from the CDI dataset; no measurement error or data revisions are discussed.
  • domain assumption Bootstrap resampling of time blocks is valid for this small panel of states.
    Stratified block bootstrap is standard, but its validity with 51 states and correlated spatial units is not examined in the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Bagging Ensemble Regression with Data Integration for Time Series-Based Diabetes Prediction." pith.science (2026). https://pith.science/paper/DAX4AM3O

@misc{pith2026250613786,
  author       = {Pith},
  title        = {Pith review of: Enhancing Bagging Ensemble Regression with Data Integration for Time Series-Based Diabetes Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DAX4AM3O}},
  note         = {Machine review of arXiv:2506.13786}
}
read the original abstract

Diabetes is a chronic metabolic disease characterized by elevated blood glucose levels, leading to complications like heart disease, kidney failure, and nerve damage. Accurate state-level predictions are vital for effective healthcare planning and targeted interventions, but in many cases, data for necessary analyses are incomplete. This study begins with a data engineering process to integrate diabetes-related datasets from 2011 to 2021 to create a comprehensive feature set. We then introduce an enhanced bagging ensemble regression model (EBMBag+) for time series forecasting to predict diabetes prevalence across U.S. cities. Several baseline models, including SVMReg, BDTree, LSBoost, NN, LSTM, and ERMBag, were evaluated for comparison with our EBMBag+ algorithm. The experimental results demonstrate that EBMBag+ achieved the best performance, with an MAE of 0.41, RMSE of 0.53, MAPE of 4.01, and an R2 of 0.9.

Figures

Figures reproduced from arXiv: 2506.13786 by the authors.

Figure 1
Figure 1. Average Percentage for Diabetes, Alcohol Consumption, Kidney Disease, [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. System Architecture: Data Engineering & Predictive Algorithms [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Average prediction performance of models [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 24 canonical work pages

  1. [1]

    ADS Authors: Statistics about diabetes.https://diabetes.org/ about-diabetes/statistics/about-diabetes(2024), [Accessed 01-October- 2024]

    work page 2024

  2. [2]

    Awad, M., Khanna, R.: Support Vector Regression, pp. 67–80. Apress, Berkeley, CA (2015)

    work page 2015

  3. [3]

    BEA authors: Personal income by state.https://www.bea.gov/data/ income-saving/personal-income-by-state(2024), [Accessed 01-October-2024]

    work page 2024

  4. [4]

    BLS Authors: Local area unemployment statistics.https://www.bls.gov/lau/ tables.htm#mstate(2024), [Accessed 01-October-2024]

    work page 2024

  5. [5]

    gov/diabetes/php/data-research/index.html(2024), [Accessed on 15-May- 2024]

    CDC-Diabetes Authors: National diabetes statistics report.https://www.cdc. gov/diabetes/php/data-research/index.html(2024), [Accessed on 15-May- 2024]

    work page 2024

  6. [6]

    CDC Population Health Authors: U.S. Chronic Disease Indicators (CDI), 2023 Release.https://data.cdc.gov/Chronic-Disease-Indicators/ U-S-Chronic-Disease-Indicators-CDI-2023-Release/g4ie-h725/about_data (2023), [Accessed 01-June-2024]

    work page 2023

  7. [7]

    PeerJ Computer Science7(e623) (2021)

    Chicco, D., Warrens, M.J., Jurman, G.: The coefficient of determination r-squared is more informative than smape, mae, mape, mse and rmse in regression analysis evaluation. PeerJ Computer Science7(e623) (2021). https://doi.org/10.7717/peerj-cs.623

    doi:10.7717/peerj-cs.623 2021

  8. [8]

    Healthcare Analytics5, 100301 (2024)

    Dharmarathne, G., Jayasinghe, T.N., Bogahawaththa, M., Meddage, D., Rath- nayake, U.: A novel machine learning approach for diagnosing diabetes with a self-explainable interface. Healthcare Analytics5, 100301 (2024). https://doi.org/https://doi.org/10.1016/j.health.2024.100301

    arXiv 2024

Show all 31 references
  1. [9]

    International Journal of Intelligent Systems 2024(1), 6688934 (2024)

    El-Sofany, H., El-Seoud, S.A., Karam, O.H., Abd El-Latif, Y.M., Taj-Eddin, I.A.T.F.: A proposed technique using machine learning for the prediction of di- abetes disease through a mobile app. International Journal of Intelligent Systems 2024(1), 6688934 (2024). https://doi.org...

    2024 doi

  2. [10]

    An- nals of Statistics pp

    Friedman, J.H.: Greedy function approximation: a gradient boosting machine. An- nals of Statistics pp. 1189–1232 (2001)

    2001

  3. [11]

    Frontiers in Genetics14(2023)

    Ganie, S.M., Pramanik, P.K.D., Bashir Malik, M., Mallik, S., Qin, H.: An ensemble learning approach for diabetes prediction using boosting techniques. Frontiers in Genetics14(2023). https://doi.org/10.3389/fgene.2023.1252159 Predicting Diabetes on Time Series 15

    2023

  4. [12]

    Applied Soft Computing108, 107483 (2021)

    He, L., Madathil, S.C., Servis, G., Khasawneh, M.T.: Neural network-based multi- task learning for inpatient flow classification and length of stay prediction. Applied Soft Computing108, 107483 (2021)

    2021

  5. [13]

    Geoscientific Model Development15(14), 5481–5487 (2022)

    Hodson, T.O.: Root-mean-square error (rmse) or mean absolute error (mae): when to use them or not. Geoscientific Model Development15(14), 5481–5487 (2022). https://doi.org/10.5194/gmd-15-5481-2022

    2022 doi

  6. [14]

    Diagnostics13(14) (2023)

    Iparraguirre-Villanueva, O., Espinola-Linares, K., Flores Casta˜ neda, R.O., Cabanillas-Carbonell, M.: Application of machine learning models for early de- tection and accurate classification of type 2 diabetes. Diagnostics13(14) (2023). https://doi.org/10.3390/diagnostics13142383

    2023 doi

  7. [15]

    ICT Express7(4), 432–439 (2021)

    Khanam, J.J., Foo, S.Y.: A comparison of machine learning algo- rithms for diabetes prediction. ICT Express7(4), 432–439 (2021). https://doi.org/https://doi.org/10.1016/j.icte.2021.02.004

    2021 doi

  8. [16]

    Scientific Reports 13(1), 21044 (2023)

    Lim, H., Kim, G., Choi, J.H.: Advancing diabetes prediction with a progressive self-transfer learning framework for discrete time series data. Scientific Reports 13(1), 21044 (2023). https://doi.org/10.1038/s41598-023-48463-0

    2023 doi

  9. [17]

    Scientific Reports14(1), 2102 (2024)

    Lugner, M., Rawshani, A., Helleryd, E., Eliasson, B.: Identifying top ten predictors of type 2 diabetes through machine learning analysis of uk biobank data. Scientific Reports14(1), 2102 (2024). https://doi.org/10.1038/s41598-024-52023-5

    2024 doi

  10. [18]

    Expert Systems with Applications177, 114905 (2021)

    Men, L., Ilk, N., Tang, X., Liu, Y.: Multi-disease prediction using lstm recurrent neural networks. Expert Systems with Applications177, 114905 (2021)

    2021

  11. [19]

    Multimedia Tools and Applications83(13), 38523–38549 (2024)

    Modak, S.K.S., Jha, V.K.: Diabetes prediction model using machine learn- ing techniques. Multimedia Tools and Applications83(13), 38523–38549 (2024). https://doi.org/10.1007/s11042-023-16745-4

    2024 doi

  12. [20]

    Neurocomputing192, 38–48 (jun 2016)

    de Myttenaere, A., Golden, B., Le Grand, B., Rossi, F.: Mean absolute per- centage error for regression models. Neurocomputing192, 38–48 (jun 2016). https://doi.org/10.1016/j.neucom.2015.12.114

    2016 doi

  13. [21]

    National Cancer Institue Authors: Census Tract Population Data Dictio- nary.https://seer.cancer.gov/censustract-pops/popdictract.html(2024), [Accessed 01-June-2024]

    2024

  14. [22]

    The Computer Jour- nal62(3), 394–413 (2019)

    Scriney, M., McCarthy, S., McCarren, A., Cappellari, P., , Roantree, M.: Automat- ing data mart construction from semi-structured data sources. The Computer Jour- nal62(3), 394–413 (2019). https://doi.org/doi.org/10.1093/comjnl/bxy064

    2019 doi

  15. [23]

    In: 2023 IEEE In- ternational Conference on Big Data (BigData)

    Scriney, M., Timilsina, M., Curry, E., Porwol, L., Nie, D., Dahley, D., Fernandez, J.B., D’Aquin, M., Roantree, M.: Engineering data assets for public health applications: A covid-19 case study. In: 2023 IEEE In- ternational Conference on Big Data (BigData). pp. 1853–1862 (202...

    2023

  16. [24]

    International Breastfeeding Journal16(1), 1–13 (2021)

    Silva, M.D.B., de Oliveira, R.D.V.C., da Alves, S.B.D., Melo, E.C.P.: Predict- ing risk of early discontinuation of exclusive breastfeeding at a brazilian referral hospital for high-risk neonates and infants: A decision-tree analysis. International Breastfeeding Journal16(1), ...

    2021

  17. [25]

    Healthcare Technology Letters10(1-2), 1–10 (2023)

    Tasin, I., Nabil, T.U., Islam, S., Khan, R.: Diabetes prediction using machine learning and explainable ai techniques. Healthcare Technology Letters10(1-2), 1–10 (2023). https://doi.org/https://doi.org/10.1049/htl2.12039

    2023 doi

  18. [26]

    U.S. Census: Explore census data 2010 & 2020.https://data.census.gov/table/ DECENNIALPL2020.H1?t=Housing:Housing%20Units&d=DEC%20Redistricting% 20Data%20(PL%2094-171)(2024), [Accessed on 01-May-2024]

    2024

  19. [27]

    Census Authors: American community survey.https://data.census.gov/ advanced(2024), [Accessed 01-October-2024] 16 Ngo, V.M

    U.S. Census Authors: American community survey.https://data.census.gov/ advanced(2024), [Accessed 01-October-2024] 16 Ngo, V.M. et al

    2024

  20. [28]

    Census Authors: State population by characteristics: 2020- 2023.https://www.census.gov/data/tables/time-series/demo/popest/ 2020s-state-detail.html(2024), [Accessed 01-June-2024]

    U.S. Census Authors: State population by characteristics: 2020- 2023.https://www.census.gov/data/tables/time-series/demo/popest/ 2020s-state-detail.html(2024), [Accessed 01-June-2024]

    2024

  21. [29]

    WHO Authors: Diabetes overview.https://www.who.int/health-topics/ diabetes#tab=tab_1(2024), [Accessed on Oct 1st, 2024]

    2024

  22. [30]

    Climate Research30, 79–82 (2005)

    Willmott, C.J., Matsuura, K.: Advantages of the mean absolute error (mae) over the root mean square error (rmse) in assessing average model performance. Climate Research30, 79–82 (2005). https://doi.org/10.3354/cr030079

    2005 doi

  23. [31]

    IEEE Transactions on Artificial Intelligence5, 1728 (2023)

    Zhao, C., Peng, R., Wu, D.: Bagging and boosting fine-tuning for ensemble learn- ing. IEEE Transactions on Artificial Intelligence5, 1728 (2023)

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.