INTERPRETABLE MACHINE LEARNING FOR DENGUE RISK STRATIFICATION: INTEGRATING CLINICAL PROFILES, DIAGNOSTIC MARKERS, AND CROSS-MODEL PREDICTOR SIGNATURES

Authors

DOI:

https://doi.org/10.67772/qxhwcz89

Keywords:

Dengue, interpretable machine learning, risk stratification, Random Forest, Logistic Regression

Abstract

Despite the substantial clinical heterogeneity of dengue infection, its early recognition can still be difficult, particularly when its clinical presentation overlaps with other febrile illnesses, which are also borne by mosquitoes. Interpretable prediction models can be used to facilitate the process of risk stratification by combining routinely available clinical and diagnostic data. This study aimed to compare Logistic Regression and Random Forest models for dengue risk stratification and evaluate their discrimination, calibration, robustness, and cross-model predictor importance. A retrospective secondary-data analysis included 299 participants with valid dengue outcomes, comprising 56 dengue-positive and 243 dengue-negative cases. Following data-quality screening, 56 predictors were retained. Models were developed using a stratified training–test design, optimized through cross-validation, and evaluated using ROC-AUC, PR-AUC, bootstrap confidence intervals, repeated cross-validation, calibration measures, and permutation importance. Logistic Regression and Random Forest achieved held-out ROC-AUCs of 0.9443 and 0.9425 and PR-AUCs of 0.7947 and 0.7606, respectively. Repeated cross-validation produced mean ROC-AUCs of 0.7148 and 0.7525. Random Forest demonstrated better calibration, with a lower Brier score (0.1334 vs. 0.1697) and expected calibration error (0.1977 vs. 0.2818). Common influential factors were health centre, thick blood smear, weight, joint pain, rapid diagnostic testing, muscle pain, and hepatomegaly. The models were both helpful for discrimination, and Random Forest demonstrated better repeated-validation ROC-AUC and calibration. There is a need for external validation for wider clinical use.

References

1.Baharom, M., Ahmad, N., Hod, R., & Abdul Manaf, M. R. (2022). Dengue early warning system as outbreak prediction tool: a systematic review. Risk Management and Healthcare Policy, 871-886.

2.Castro, M. C., Wilson, M. E., & Bloom, D. E. (2017). Disease and economic burdens of dengue. The Lancet Infectious Diseases, 17(3), e70-e78.

3.Darshan, B. B., Holla, R., Unnikrishnan, B., Achappa, B., Poovattil, R., Sharma, A., ... & Prateek, S. (2024). Clinical and seasonal pattern of dengue in a tertiary care hospital of South West India. F1000Research, 12, 817.

4.Flórez, J. E. S., Velasquez, K. M., Cardona, Á. M. S., Jaramillo, B. N. R., Díaz, Y. E. O., Cardona, L. S. G., & Naranjo, M. A. (2024). Clinical manifestations of dengue in children and adults in a hyperendemic region of Colombia. The American journal of tropical medicine and hygiene, 110(5), 971.

5.Goweda, R., & Faisal, A. (2020). A study of clinical features and laboratory profile of dengue fever in outpatient setting. Malaysian Journal of Public Health Medicine, 20(2), 94-100.

6.Gui, H., Gwee, S., Koh, J., & Pang, J. (2021). Weather factors associated with reduced risk of dengue transmission in an urbanized tropical city. International journal of environmental research and public health, 19(1), 339.

7.Guo, P., Liu, T., Zhang, Q., Wang, L., Xiao, J., Zhang, Q., ... & Ma, W. (2017). Developing a dengue forecast model using machine learning: A case study in China. PLoS neglected tropical diseases, 11(10), e0005973.

8.Halstead, S. (2019). Recent advances in understanding dengue. F1000Research, 8, F1000-Faculty.

9.Ho, T. S., Weng, T. C., Wang, J. D., Han, H. C., Cheng, H. C., Yang, C. C., ... & Liu, C. C. (2020). Comparing machine learning with case-control models to identify confirmed dengue cases. PLoS neglected tropical diseases, 14(11), e0008843.

10.Khan, M. B., Yang, Z. S., Lin, C. Y., Hsu, M. C., Urbina, A. N., Assavalapsakul, W., ... & Wang, S. F. (2023). Dengue overview: An updated systemic review. Journal of infection and public health, 16(10), 1625-1642.

11.Madewell, Z. J., Rodriguez, D. M., Thayer, M. B., Rivera-Amill, V., Paz-Bailey, G., Adams, L. E., & Wong, J. M. (2025). Machine learning for predicting severe dengue in Puerto Rico. Infectious diseases of poverty, 14(1), 5.

12.Muller, D. A., Depelsenaire, A. C., & Young, P. R. (2017). Clinical and laboratory diagnosis of dengue virus infection. The Journal of infectious diseases, 215(suppl_2), S89-S95.

13.Nakase, T., Giovanetti, M., Obolski, U., & Lourenço, J. (2024). Population at risk of dengue virus transmission has increased due to coupled climate factors and population growth. Communications Earth & Environment, 5(1), 475.

14.Ong, J., Aik, J., & Ng, L. C. (2021). Adult Aedes abundance and risk of dengue transmission. PLoS Neglected Tropical Diseases, 15(6), e0009475.

15.Ouedraogo, I., Some, B., & Ouedraogo, I. (2025). Tabular dataset for AI-based vector-borne disease prediction (Version 2) [Data set]. Mendeley Data. https://doi.org/10.17632/cf49v47z4c.2

16.Patel, J. P., Saiyed, F., & Hardaswani, D. (2024). Dengue fever accompanied by neurological manifestations: challenges and treatment. Cureus, 16(5).

17.Paz-Bailey, G., Adams, L. E., Deen, J., Anderson, K. B., & Katzelnick, L. C. (2024). Dengue. The Lancet, 403(10427), 667-682.

18.Siriyasatien, P., Chadsuthi, S., Jampachaisri, K., & Kesorn, K. (2018). Dengue epidemics prediction: A survey of the state-of-the-art based on data science processes. IEEE Access, 6, 53757-53795.

19.Trivedi, S., & Chakravarty, A. (2022). Neurological complications of dengue fever. Current neurology and neuroscience reports, 22(8), 515-529.

20.Tuhin, I. A., Siam, A. F. K., Shanto, M. M. R., Mia, M. R., Mahmud, I., & Ghosh, A. (2025). An interpretable machine learning model for dengue detection with clinical hematological data. Healthcare Analytics, 100430.

21.Watts, M. J., Kotsila, P., Mortyn, P. G., Sarto i Monteys, V., & Urzi Brancati, C. (2020). Influence of socio-economic, demographic and climate factors on the regional distribution of dengue in the United States and Mexico. International journal of health geographics, 19(1), 44.

Wilder-Smith, A., Ooi, E. E., Horstick, O., & Wills, B. (2019). Dengue. The Lancet, 393(10169), 350-363.

Downloads

Published

2026-10-28