arxiv: 2604.18642 · v1 · submitted 2026-04-19 · 🧮 math.GM

Recognition: unknown

Climate-Driven Dengue Forecasting in Bangladesh: Division-Specific Feature-Set Design and Lag Structure

Faizunnesa Khondaker, Md. Kamrujjaman

Authors on Pith no claims yet

Pith reviewed 2026-05-10 05:45 UTC · model grok-4.3

classification 🧮 math.GM

keywords dengue forecastingclimate variableslag analysisBangladeshfeature engineeringneural networksSARIMAXmosquito-borne disease

0 comments

The pith

Division-specific climate feature sets and lag structures yield the lowest-error dengue forecasts for Dhaka and Barishal.

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

Bangladesh shows large year-to-year swings in dengue cases that are partly shaped by weather conditions affecting mosquito breeding and virus transmission. The paper focuses on two divisions with different case burdens and tests whether careful choices in how to represent rain and sun exposure can improve monthly forecasts. Four feature sets are built by swapping between total rainfall and rainy-day counts for wetness and between sunshine hours and sun-day counts for exposure, with temperature and humidity included in every set. These sets are paired with climate lags of zero to four months and evaluated both with and without the previous month’s dengue count as an extra input. Several model families are compared on data from 2022 to 2025, and the combinations that minimize error turn out to be different for each division.

Core claim

By varying the representation of wetness and sunshine in four climate feature sets and applying consistent 0–4 month lags to monthly dengue records, the authors show that an artificial neural network paired with one particular feature set produces the lowest errors for Dhaka while a seasonal autoregressive integrated moving average model paired with a different feature set performs best for Barishal.

What carries the argument

Division-specific feature-set design that creates four variants by alternating rainfall versus rainy days and sun hours versus sun days, then combines them with 0–4 month climate lags and an optional one-month lagged dengue term.

Load-bearing premise

The chosen 0–4 month lags and hand-designed feature sets capture stable relationships with dengue incidence rather than chance patterns that appear only in the 46-month data window.

What would settle it

If the recommended feature sets and models produce markedly higher forecast errors when applied to dengue incidence recorded after October 2025, the stability of the lag structures and feature choices would be contradicted.

Figures

Figures reproduced from arXiv: 2604.18642 by Faizunnesa Khondaker, Md. Kamrujjaman.

**Figure 2.** Figure 2: Division-wise heatmap of infected population in Bangladesh in 2025, [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Figure 3: Dengue cases in Dhaka and Barishal: (a) monthly reported cases [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 4.** Figure 4: Time-series of key climate variables in the Dhaka and Barishal divi [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

**Figure 5.** Figure 5: Month-wise boxplots of key climate variables in the Dhaka and Bar [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

**Figure 6.** Figure 6: STL decomposition of monthly dengue cases for Dhaka, showing [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗

**Figure 7.** Figure 7: STL decomposition of monthly dengue cases for Barishal, showing [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗

**Figure 8.** Figure 8: ACF and PACF plots of monthly dengue cases for Dhaka. [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗

**Figure 9.** Figure 9: ACF and PACF plots of monthly dengue cases for Barishal. [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗

**Figure 10.** Figure 10: Heatmap of lagged (0-4 months) Pearson correlations between [PITH_FULL_IMAGE:figures/full_fig_p016_10.png] view at source ↗

**Figure 11.** Figure 11: Bar plot of the strongest lagged Pearson correlation for each climate [PITH_FULL_IMAGE:figures/full_fig_p017_11.png] view at source ↗

**Figure 12.** Figure 12: SARIMAX results for Dhaka division. (a) Full series observed cases [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗

**Figure 13.** Figure 13: SARIMAX results for the Dhaka division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p020_13.png] view at source ↗

**Figure 14.** Figure 14: SARIMAX results for Barishal division. (a) Full series observed [PITH_FULL_IMAGE:figures/full_fig_p022_14.png] view at source ↗

**Figure 15.** Figure 15: SARIMAX results for the Barishal division, showing the error met [PITH_FULL_IMAGE:figures/full_fig_p023_15.png] view at source ↗

**Figure 16.** Figure 16: MPR results for Dhaka division. (a) Full series observed cases with [PITH_FULL_IMAGE:figures/full_fig_p024_16.png] view at source ↗

**Figure 17.** Figure 17: MPR results for the Dhaka division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p025_17.png] view at source ↗

**Figure 18.** Figure 18: MPR results for Barishal division. (a) Full series observed cases [PITH_FULL_IMAGE:figures/full_fig_p026_18.png] view at source ↗

**Figure 19.** Figure 19: MPR results for the Barishal division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p027_19.png] view at source ↗

**Figure 20.** Figure 20: ANN results for Dhaka division. (a) Full series observed cases with [PITH_FULL_IMAGE:figures/full_fig_p028_20.png] view at source ↗

**Figure 21.** Figure 21: ANN results for the Dhaka division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p029_21.png] view at source ↗

**Figure 22.** Figure 22: ANN results for Barishal division. (a) Full series observed cases [PITH_FULL_IMAGE:figures/full_fig_p030_22.png] view at source ↗

**Figure 23.** Figure 23: ANN results for the Barishal division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p031_23.png] view at source ↗

**Figure 24.** Figure 24: XGBoost results for Dhaka division. (a) Full series observed cases [PITH_FULL_IMAGE:figures/full_fig_p032_24.png] view at source ↗

**Figure 25.** Figure 25: XGBoost results for the Dhaka division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p033_25.png] view at source ↗

**Figure 26.** Figure 26: XGBoost results for Barishal division. (a) Full series observed cases [PITH_FULL_IMAGE:figures/full_fig_p034_26.png] view at source ↗

**Figure 27.** Figure 27: XGBoost results for the Barishal division, showing the error metrices [PITH_FULL_IMAGE:figures/full_fig_p035_27.png] view at source ↗

**Figure 28.** Figure 28: Best-performing models by feature set for Dhaka division (a) Full [PITH_FULL_IMAGE:figures/full_fig_p036_28.png] view at source ↗

**Figure 29.** Figure 29: Best-performing models by feature set for Barishal division (a) Full [PITH_FULL_IMAGE:figures/full_fig_p037_29.png] view at source ↗

read the original abstract

Bangladesh exhibits marked year-to-year variability in dengue, partly driven by meteorological fluctuations that shape \textit{Aedes} breeding-site persistence, mosquito development, and transmission. We exploit a contrast between Dhaka (consistently high burden) and Barishal (recently rising burden despite lower population density) and frame feature-set design and predictor structure as the main methodological contributions. Using monthly dengue data from DGHS \cite{DGHS} and meteorological data from World Weather Online \cite{Weather} for January 2022--October 2025, we compare four climate feature sets that vary wetness (rainy days vs.\ rainfall) and sunshine (sun days vs.\ sun hours), while temperature and humidity appear in all sets. We evaluate two predictor configurations: lagged climate covariates only, and lagged climate covariates plus 1-month lagged dengue incidence ($Y_{t-1}$). Climate lags (0--4 months) are applied in correlation and forecasting. Both divisions show similar delayed associations: rainfall metrics peak positively near a 2-month lag, humidity near a 1-month lag, sunshine metrics are most negative around a 2-month lag, and temperature is weakly positive at longer lags. We then benchmark MPR, ANN, XGBoost, and SARIMAX across all sets. Best performance differs: Dhaka favors ANN-1 with SET-1 (RMSE=2176.70, MAE=1282.00, MAPE=31.54\%), whereas Barishal favors SARIMAX(0,1,1)(1,0,0,12) with SET-2 (RMSE=817.56, MAE=717.78, MAPE=39.96\%). Analyses use consistent monthly aggregation and division-specific tuning.

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.

Desk Editor's Note private letter to a colleague

read the letter

The paper applies routine ML and time-series models to a short dengue series in two Bangladesh divisions and reports division-specific winners, but the 46-month length plus in-sample lag and feature tuning make the performance claims unreliable. It compares four hand-designed climate feature sets (varying rainy days versus total rainfall and sun days versus hours) plus lagged incidence, then benchmarks MPR, ANN, XGBoost, and SARIMAX on monthly data from 2022–2025. The lag correlations line up with known mosquito biology—rainfall and sunshine effects strongest around two months, humidity near one month—which is a small but useful check. The division contrast itself is reasonable: Dhaka has steady high burden while Barishal has been rising, so showing that no single model or feature set wins everywhere is a practical observation. The concrete RMSE, MAE, and MAPE numbers are reported clearly, and the authors note that temperature and humidity appear in every set while wetness and sunshine metrics are swapped. That level of transparency on the feature variants is better than many applied forecasting papers. The main weakness is the data scale and validation. Forty-six monthly points is thin for seasonal SARIMAX or flexible ANNs, and choosing lags 0–4 by inspecting correlations on the same series used for fitting introduces circularity. The abstract gives no description of train-test splits, rolling-origin cross-validation, or blocked temporal hold-outs, so the quoted “best” models for each division could simply be fitting noise. With multiple models, four feature sets, and division-specific tuning all evaluated on the same short window, the performance gap between ANN-1 for Dhaka and SARIMAX for Barishal is hard to trust as stable. This is useful reading for public-health teams in Bangladesh or nearby countries who need local dengue forecasts and are willing to adapt the feature ideas. Readers focused on general climate-disease methods will find little new, but the applied contrast between divisions gives it some value. It deserves peer review because the data source and question are real and the lag patterns are biologically plausible, even though the evaluation needs a proper temporal validation scheme added before the results can be taken as reliable forecasts.

Referee Report

3 major / 2 minor

Summary. The manuscript claims that division-specific climate feature sets (varying wetness and sunshine metrics, with temperature and humidity common) and lag structures (0-4 months selected via correlation) enable improved dengue forecasting in Bangladesh. Using 46 monthly observations (Jan 2022-Oct 2025) from DGHS and World Weather Online, it benchmarks MPR, ANN, XGBoost, and SARIMAX models and reports that Dhaka performs best with ANN-1 + SET-1 (RMSE=2176.70, MAE=1282.00, MAPE=31.54%) while Barishal performs best with SARIMAX(0,1,1)(1,0,0,12) + SET-2 (RMSE=817.56, MAE=717.78, MAPE=39.96%), with similar delayed climate associations in both divisions.

Significance. If the performance differences hold under proper temporal validation, the work could contribute to localized dengue forecasting by showing that feature-set design and model choice vary meaningfully across divisions. The explicit reporting of lag correlations (rainfall positive at ~2 months, humidity at ~1 month, sunshine negative at ~2 months) and concrete error metrics for each model-feature combination provides transparency and a basis for replication. However, the short series and tuning process limit generalizability claims.

major comments (3)

[Abstract] Abstract: The best-model claims rest on RMSE/MAE/MAPE values obtained after selecting lags 0-4 and tuning four hand-designed feature sets plus model hyperparameters on the identical 46-month series; no train/test split, rolling-origin CV, or blocked temporal validation is described, so the quoted metrics (e.g., Dhaka ANN-1 RMSE=2176.70) are not demonstrably out-of-sample.
[Abstract] Abstract (lag and feature selection): Correlation-based lag choice and division-specific feature-set tuning are performed on the full data later used for fitting; this circularity means the reported performance advantage of ANN-1/SET-1 over other combinations in Dhaka (and SARIMAX/SET-2 in Barishal) may reflect capitalization on noise rather than stable climate-dengue relationships.
[Abstract] Abstract (model comparison): With n≈46 and flexible models (ANN, XGBoost, seasonal SARIMAX) plus multiple feature variants, the assertion that 'best performance differs' between divisions lacks supporting statistical tests for error differences or evidence that the division-specific tuning does not overfit the limited sample.

minor comments (2)

[Abstract] The abstract states the period as January 2022--October 2025 but does not explicitly note the resulting 46 observations; adding this would help readers immediately gauge sample-size constraints.
[Abstract] Abbreviations such as 'ANN-1', 'SET-1', and the exact composition of the four feature sets are used in the abstract without prior definition, forcing readers to consult later sections for interpretation.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the careful and constructive review. The comments correctly identify key limitations in validation and the risks of overfitting with a short time series. We address each point below and commit to revisions that incorporate proper temporal validation, separated selection procedures, and statistical comparisons while preserving the exploratory value of the division-specific analysis.

read point-by-point responses

Referee: The best-model claims rest on RMSE/MAE/MAPE values obtained after selecting lags 0-4 and tuning four hand-designed feature sets plus model hyperparameters on the identical 46-month series; no train/test split, rolling-origin CV, or blocked temporal validation is described, so the quoted metrics (e.g., Dhaka ANN-1 RMSE=2176.70) are not demonstrably out-of-sample.

Authors: We agree that the reported metrics are in-sample because no explicit train/test split or rolling-origin CV is described in the current manuscript. With only 46 observations, further splitting reduces statistical power, but this does not justify omitting validation. In the revised manuscript we will add a rolling-origin cross-validation procedure (using the first 34 months for training and expanding the test window) and report out-of-sample RMSE/MAE/MAPE for the final selected models. This will allow readers to evaluate whether the performance differences persist under temporal validation. revision: yes
Referee: Correlation-based lag choice and division-specific feature-set tuning are performed on the full data later used for fitting; this circularity means the reported performance advantage of ANN-1/SET-1 over other combinations in Dhaka (and SARIMAX/SET-2 in Barishal) may reflect capitalization on noise rather than stable climate-dengue relationships.

Authors: The referee accurately notes the circularity. Lag selection (0-4 months) and feature-set design were performed on the full series to identify biologically plausible structures supported by prior dengue literature. To correct this, the revision will separate the processes: lag and feature selection will be conducted on an initial training window only, with model evaluation performed on a subsequent held-out period. We will also report the raw correlation coefficients for each lag and division independently of the forecasting results. revision: yes
Referee: With n≈46 and flexible models (ANN, XGBoost, seasonal SARIMAX) plus multiple feature variants, the assertion that 'best performance differs' between divisions lacks supporting statistical tests for error differences or evidence that the division-specific tuning does not overfit the limited sample.

Authors: We accept that formal statistical tests are needed to support claims of differential performance across divisions. The revision will include Diebold-Mariano tests for pairwise forecast accuracy comparisons and bootstrap confidence intervals around the error metrics. We will also add an explicit limitations paragraph discussing the small sample size, the exploratory character of the division-specific tuning, and the fact that similar lag patterns (rainfall ~2 months, humidity ~1 month) appear in both divisions, providing partial corroboration despite the limited n. revision: yes

Circularity Check

1 steps flagged

Lag selection via correlation inspection and division-specific model tuning on the full 46-month series makes reported 'best performance' in-sample fitted results

specific steps

fitted input called prediction [Abstract (lag application and benchmarking paragraphs)]
"Climate lags (0--4 months) are applied in correlation and forecasting. ... We then benchmark MPR, ANN, XGBoost, and SARIMAX across all sets. Best performance differs: Dhaka favors ANN-1 with SET-1 (RMSE=2176.70, MAE=1282.00, MAPE=31.54%), whereas Barishal favors SARIMAX(0,1,1)(1,0,0,12) with SET-2 (RMSE=817.56, MAE=717.78, MAPE=39.96%)."

Lags are selected by inspecting correlations on the identical 46-month series later used for model fitting and evaluation; feature sets are hand-designed and models are chosen by minimizing the same performance metrics on this data, so the reported 'best performance' values are statistically forced by the selection process rather than out-of-sample forecasts.

full rationale

The paper's central claim identifies division-specific optimal models and feature sets via RMSE/MAE/MAPE on the Jan 2022–Oct 2025 data. Lags 0–4 are chosen after correlation analysis on that same series, four feature sets are hand-designed, and four model families are benchmarked with per-division tuning. No train/test split, rolling CV, or blocked validation is described in the abstract, so the quoted performance numbers reduce to results of selection and fitting on the target observations rather than independent predictions. This matches the 'fitted input called prediction' pattern exactly. No self-citation load-bearing, ansatz smuggling, or self-definitional steps appear; the derivation is otherwise self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

3 free parameters · 2 axioms · 0 invented entities

The central claim rests on data-driven lag selection and model tuning rather than external benchmarks or parameter-free derivations; no new physical entities are introduced.

free parameters (3)

Climate lag window
0-4 month lags selected after inspecting correlations on the study data.
Feature-set variants
Four combinations of wetness and sunshine metrics chosen by the authors.
Model hyperparameters
ANN architecture, SARIMAX orders, and XGBoost parameters tuned to the same series.

axioms (2)

domain assumption Monthly aggregation preserves the relevant transmission dynamics
Stated in the methods description of consistent monthly aggregation.
domain assumption Lagged climate variables are exogenous to dengue incidence
Implicit in the use of climate as predictors without feedback terms.

pith-pipeline@v0.9.0 · 5630 in / 1512 out tokens · 58774 ms · 2026-05-10T05:45:25.854016+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

30 extracted references · 19 canonical work pages

[1]

Daily Dengue Status Report

Directorate General of Health Services [DGHS]. Daily Dengue Status Report. Bangladesh Ministry of Health and Family Welfare. Retrieved November 10, 2025, fromhttps://old.dghs.gov.bd/index.php/bd/home/ 5200-daily-dengue-status-report

2025
[2]

Bangladesh weather

World Weather Online. Bangladesh weather. InWorld Weather Online. Re- trieved November 21, 2025, fromhttps://www.worldweatheronline.com/ bangladesh-weather.aspx

2025
[3]

H., Nam, K

Lee, S. H., Nam, K. W., Jeong, J. Y., Yoo, S. J., Koh, Y. S., Lee, S., ... & Lee, K. H. (2013). The effects of climate change and globalization on mosquito vectors: evidence from Jeju Island, South Korea on the potential for Asian tiger mosquito (Aedes albopictus) influxes and survival from Vietnam rather than Japan.PloS one, 8(7), e68512.https://doi.org/...

work page doi:10.1371/journal.pone.0068512 2013
[4]

Simo, F. B. N., Bigna, J. J., Kenmoe, S., Ndangang, M. S., Temfack, E., Moundipa, P. F., & Demanou, M. (2019). Dengue virus infection in people residing in Africa: a systematic review and meta-analysis of prevalence studies.Scientific reports, 9(1), 13626.https://doi.org/10.1038/s41598-019-50135-x

work page doi:10.1038/s41598-019-50135-x 2019
[5]

Ong, J., Aik, J., and Ng, L. C. (2021). Short report: AdultAedesabundance and risk of dengue transmission.PLoS Neglected Tropical Diseases, 15(6), e0009475.https: //doi.org/10.1371/journal.pntd.0009475

work page doi:10.1371/journal.pntd.0009475 2021
[6]

Murray, N. E. A., Quam, M. B., and Wilder-Smith, A. (2013). Epidemiology of dengue: Past, present and future prospects.Clinical Epidemiology, 5, 299–309. https://doi.org/10.2147/CLEP.S34440 40

work page doi:10.2147/clep.s34440 2013
[7]

(2024, April 16).Life cy- cle of Aedes mosquitoes.https://www.cdc.gov/mosquitoes/about/ life-cycle-of-aedes-mosquitoes.html

Centers for Disease Control and Prevention. (2024, April 16).Life cy- cle of Aedes mosquitoes.https://www.cdc.gov/mosquitoes/about/ life-cycle-of-aedes-mosquitoes.html

2024
[8]

Nasirian, H. (2025). Factors influencing establishment of dengue fever vectors in urban areas.East Mediterr Health J, 31(3), 163-165

2025
[9]

Haider, M., & Turner, J. (2015). Variables that may affect the transmission of dengue–a case study for health management in Asia.Topics in Public Health

2015
[10]

E., Hossain, S., and Hanesiak, J

Islam, S., Haque, C. E., Hossain, S., and Hanesiak, J. (2021). Climate variability, dengue vector abundance and dengue fever cases in Dhaka, Bangladesh: a time-series study.Atmosphere, 12, 905.https://doi.org/10.3390/atmos12070905

work page doi:10.3390/atmos12070905 2021
[11]

P., Zhou, W., Ren, C., Marshall, J., & Yuan, H

Hossain, M. P., Zhou, W., Ren, C., Marshall, J., & Yuan, H. Y. (2022). Prediction of dengue annual incidence using seasonal climate variability in Bangladesh between 2000 and 2018.PLOS Global Public Health, 2(5), e0000047.https://doi.org/10. 1371/journal.pgph.0000047

2022
[12]

M., Hasan, M

Hossain, S., Islam, M. M., Hasan, M. A., Chowdhury, P. B., Easty, I. A., Tusar, M. K., ... & Bashar, K. (2023). Association of climate factors with dengue incidence in Bangladesh, Dhaka City: A count regression approach.Heliyon, 9(5).https: //doi.org/10.1016/j.heliyon.2023.e16053

work page doi:10.1016/j.heliyon.2023.e16053 2023
[13]

R., Sarker, I., & Bashar, K

Karim, M. R., Sarker, I., & Bashar, K. (2023). Changing patterns of climatic risk factors for the transmission of dengue fever in Bangladesh: count model ap- proach.Journal of Public Health, 33(8), 1607-1616.https://doi.org/10.1007/ s10389-023-02137-3

2023
[14]

T., Stephenson, P., Islam, M

Muurlink, O. T., Stephenson, P., Islam, M. Z., & Taylor-Robinson, A. W. (2018). Long-term predictors of dengue outbreaks in Bangladesh: A data mining approach. Infectious Disease Modelling, 3, 322–330.https://doi.org/10.1016/j.idm.2018. 11.004

work page doi:10.1016/j.idm.2018 2018
[15]

S., & Kamrujjaman, M

Islam, S., Islam, M. S., & Kamrujjaman, M. (2025). Effectiveness of bed nets and media awareness in dengue control: A fuzzy analysis.Results in Control and Opti- mization,20, 100593.https://doi.org/10.1016/j.rico.2025.100593

work page doi:10.1016/j.rico.2025.100593 2025
[16]

Okamura and S

Dey, S. K., Rahman, M. M., Howlader, A., Siddiqi, U. R., Uddin, K. M. M., Borhan, R., & Rahman, E. U. (2022). Prediction of dengue incidents using hospitalized pa- tients, metrological and socio-economic data in Bangladesh: A machine learning approach.PloS one, 17(7), e0270933.https://doi.org/10.1371/journal.pone. 0270933 41

work page doi:10.1371/journal.pone 2022
[17]

A., Kamrujjaman, M., Molla, M

Mobin, M. A., Kamrujjaman, M., Molla, M. M., & Chen, S. (2024). Analysis of a data driven vector-borne dengue transmission model for a tropical environment in Bangladesh.International Journal of Differential Equations,2024, Article 2959770. https://doi.org/10.1155/2024/2959770

work page doi:10.1155/2024/2959770 2024
[18]

E., Ahmed, M

Alam, K. E., Ahmed, M. J., Chalise, R., Rahman, M. A., Mathin, T. T., Bhuiyan, M. I. H., ... & Hossain, D. (2025). Time series analysis of dengue incidence and its association with meteorological risk factors in Bangladesh.Plos one, 20(8), e0323238. https://doi.org/10.1371/journal.pone.0323238

work page doi:10.1371/journal.pone.0323238 2025
[19]

Chowdhury, A. H. (2025). Comparison of deep learning and gradient boosting: ANN Versus XGBoost for climate-based dengue prediction in Bangladesh.Health Science Reports, 8(4), e70714.https://doi.org/10.1002/hsr2.70714

work page doi:10.1002/hsr2.70714 2025
[20]

T., Hossain, M

Akter, T., Islam, M. T., Hossain, M. F., & Ullah, M. S. (2024). A comparative study between time series and machine learning technique to predict dengue fever in Dhaka city.Discrete Dynamics in Nature and Society, 2024(1), 2757381.https: //doi.org/10.1155/2024/2757381

work page doi:10.1155/2024/2757381 2024
[21]

J., Sultana, N., Das, A., Jui, F

Hossain, M. J., Sultana, N., Das, A., Jui, F. N., Islam, M. K., Rahman, M. M., & Rahman, M. M. (2024). Analysis of effects of meteorological variables on dengue incidence in Bangladesh using VAR and Granger causality approach.Frontiers in Public Health, 12, 1488742.https://doi.org/10.3389/fpubh.2024.1488742

work page doi:10.3389/fpubh.2024.1488742 2024
[22]

A., Fahim, H., Althobaiti, T., Alsharif, N., & Saeed, N

Azeem, M., Javaid, S., Khalil, R. A., Fahim, H., Althobaiti, T., Alsharif, N., & Saeed, N. (2023). Neural networks for the detection of COVID-19 and other diseases: prospects and challenges.Bioengineering, 10(7), 850.https://doi.org/10.3390/ bioengineering10070850

2023
[23]

M., Alkhamees, B

Bukhari, M. M., Alkhamees, B. F., Hussain, S., Gumaei, A., Assiri, A., & Ullah, S. S. (2021). An improved artificial neural network model for effective diabetes prediction. Complexity, 2021(1), 5525271.https://doi.org/10.1155/2021/5525271

work page doi:10.1155/2021/5525271 2021
[24]

H., & Mila, F

Noorunnahar, M., Chowdhury, A. H., & Mila, F. A. (2023). A tree based eXtreme Gradient Boosting (XGBoost) machine learning model to forecast the annual rice production in Bangladesh.PloS one, 18(3), e0283452.https://doi.org/10.1371/ journal.pone.0283452

2023
[25]

Guan, X., Du, Y., Ma, R., Teng, N., Ou, S., Zhao, H., & Li, X. (2023). Construction of the XGBoost model for early lung cancer prediction based on metabolic indices. BMC medical informatics and decision making, 23(1), 107.https://doi.org/10. 1186/s12911-023-02171-x 42

2023
[26]

(2022).Population & Housing Census 2022: Preliminary report (English)

Bangladesh Bureau of Statistics (BBS). (2022).Population & Housing Census 2022: Preliminary report (English). Statistics and Informatics Division, Ministry of Planning, Government of the People’s Republic of Bangladesh. Retrieved February 10, 2026, fromhttps://sid.portal.gov.bd/sites/default/files/files/sid. portal.gov.bd/publications/01ad1ffe_cfef_4811_a...

2022
[27]

Polwiang, S. (2020). The time series seasonal patterns of dengue fever and associ- ated weather variables in Bangkok (2003-2017).BMC infectious diseases, 20(1), 208. https://doi.org/10.1186/s12879-020-4902-6

work page doi:10.1186/s12879-020-4902-6 2020
[28]

Khondaker, F., Kamrujjaman, M., & Islam, M. S. (2025). Cost-effectiveness of dengue control strategies in Bangladesh: An optimal control and ACER-ICER analysis.Acta Tropica, 264, 107587.https://doi.org/10.1016/j.actatropica.2025.107587

work page doi:10.1016/j.actatropica.2025.107587 2025
[29]

Khondaker, F., & Kamrujjaman, M. (2025). Stochastic analysis of Mpox epidemi- ology with vaccination strategies and environmental persistence.Scientific Reports, 15(1), 44369.https://doi.org/10.1038/s41598-025-28135-x

work page doi:10.1038/s41598-025-28135-x 2025
[30]

Khondaker, F., Kamrujjaman, M., & Islam, M. S. (2022). Optimal control analysis of COVID-19 transmission model with physical distance and treatment.Advance in Biological Research, 3(1).https://doi.org/10.26855/abr.2022.12.001 43

work page doi:10.26855/abr.2022.12.001 2022