Climate-Driven Disease Outbreak Anomaly Detection Using PCA, DBSCAN, and Isolation Forest

Titin Wahyuni; Muhammad Faisal; Muhammad Syafa’at S. Kuba; Eny Syatriana; Muh Al Haikal

doi:10.30736/jti.v11i02.1674

Authors

Titin Wahyuni Universaitas Muhammadiyah Makassar
Muhammad Faisal Universitas Muhammadiyah Makassar
Muhammad Syafa’at S. Kuba Universitas Muhammadiyah Makassar
Eny Syatriana Universitas Muhammadiyah Makassar
Muh Al Haikal Universitas Muhammadiyah Makassar

DOI:

https://doi.org/10.30736/jti.v11i02.1674

Keywords:

Climate-driven disease, surveillance, Dangue, Anomaly Detection, Unsupervised learning, PCA, DBSCAN, Isolation Forest, Early Warning Systems

Abstract

Climate variability strongly influences dengue transmission, but prediction-focused surveillance often overlooks unusual climate-dengue configurations that emerge before or outside epidemic peaks. This study proposes a fully unsupervised anomaly-oriented framework for national-level dengue surveillance in Brazil. Weekly dengue case data were integrated with NASA POWER climate variables aggregated to the same epidemiological week scale. Seven standardized climate indicators were reduced using Principal Component Analysis (PCA); the first three components retained 93.34% of the total variance. DBSCAN was then used to identify structurally isolated climate regimes, and Isolation Forest was applied to assign anomaly severity scores and examine complementary transitional deviations. The results show that DBSCAN detected 19 epidemiological weeks (5.19%) as structural anomaly candidates, while Isolation Forest highlighted severity-ranked weeks whose case totals were not always the highest in the study period. To support interpretation, detected anomalies are evaluated using temporal identifiers, dengue case counts, climate-variable deviations, and cross-method robustness rather than treated as automatic outbreak declarations. The findings indicate that climate-driven dengue anomalies are episodic and multivariate and may occur outside peak incidence periods. The proposed framework therefore offers an explainable and reproducible screening layer for climate-informed early warning and risk prioritization in public health surveillance

Downloads

Download data is not yet available.

References

[1] R. Akhtar and A. J. McMichael, “Climate change and vector-borne diseases: An emerging global health challenge,” Environ. Res., vol. 212, p. 113195, 2022, doi: 10.1016/j.envres.2022.113195.

[2] S. Banu, W. Hu, and S. Tong, “Dengue transmission and weather variability: Advances in climate-driven modeling,” Science of the Total Environment, vol. 857, p. 159353, 2023, doi: 10.1016/j.scitotenv.2022.159353.

[3] B. M. of Health, “Dengue surveillance and epidemiological reports (DATASUS),” Ministério da Saúde, Brazil, 2023.

[4] W. H. O, “Global dengue situation update,” World Health Organization, 2023.

[5] P. A. H. Organization, “Dengue epidemiological update: Americas,” PAHO, 2023.

[6] R. Lowe, A. M. Stewart-Ibarra, D. Petrova, and et al., “Climate services for health: Predicting, preventing, and managing climate-sensitive diseases,” Science of the Total Environment, vol. 843, p. 156987, 2022, doi: 10.1016/j.scitotenv.2022.156987.

[7] T. M. Carvajal, “Machine learning methods reveal key climate factors influencing dengue outbreaks,” Sci. Rep., vol. 11, p. 3309, 2021, doi: 10.1038/s41598-021-82109-4.

[8] L. Chien and H. Yu, “Climate-driven dengue epidemic modeling,” International Journal of Infectious Diseases, vol. 118, pp. 120–130, 2022, doi: 10.1016/j.ijid.2022.02.014.

[9] D. Fisman, “Seasonality of infectious diseases: A mathematical and climate approach,” Clin. Microbiol. Rev., vol. 33, no. 4, 2020, doi: 10.1128/CMR.00100-19.

[10] S. Mak and R. Lowe, “Integrating weather data into vector-borne disease monitoring systems,” Lancet Planetary Health, vol. 5, no. 7, pp. e423–e432, 2021, doi: 10.1016/S2542-5196(21)00138-9.

[11] S. Kumar and R. Patel, “Machine learning-based detection of climate-sensitive diseases,” Sci. Rep., vol. 12, p. 1124, 2022, doi: 10.1038/s41598-022-01124-7.

[12] M. Nguyen and G. Harper, “Climate-driven outbreak prediction using multisource environmental data,” Nat. Clim. Chang., vol. 10, pp. 973–982, 2020, doi: 10.1038/s41558-020-00987-2.

[13] A. Rahman, “Dengue hotspot prediction through climate signatures,” Sci. Data, vol. 9, p. 144, 2022, doi: 10.1038/s41597-022-01234-5.

[14] R. Singh, “Seasonal climate anomalies as early warning indicators for dengue,” Science of the Total Environment, vol. 903, p. 166973, 2024, doi: 10.1016/j.scitotenv.2023.166973.

[15] D. Gómez, “PCA-enhanced multivariate climate analysis for disease forecasting,” Environmental Modelling & Software, vol. 160, p. 105513, 2023, doi: 10.1016/j.envsoft.2023.105513.

[16] P. Sharma, “Climate-driven epidemiological modeling using PCA and clustering,” Int. J. Environ. Res. Public Health, vol. 20, no. 4, p. 3201, 2023, doi: 10.3390/ijerph20043201.

[17] H. Liu and G. Chen, “Isolation Forest for anomaly detection in complex epidemiological data,” Pattern Recognit., vol. 138, p. 109402, 2023, doi: 10.1016/j.patcog.2023.109402.

[18] C. Trujillo, “Unsupervised anomaly detection for epidemiological time-series,” Artif. Intell. Med., vol. 140, p. 102546, 2023, doi: 10.1016/j.artmed.2023.102546.

[19] P. Zhang, “Hybrid clustering approaches for disease outbreak detection,” Expert Syst. Appl., vol. 185, p. 115617, 2021, doi: 10.1016/j.eswa.2021.115617.

[20] M. Thamrin et al., “Machine learning for global trade analysis: a hybrid clustering approach using DBSCAN, elbow, and SOM,” IAES International Journal of Artificial Intelligence (IJ-AI), vol. 14, no. 4, p. 3033, 2025, doi: 10.11591/ijai.v14.i4.pp3033-3046.

[21] M. M. Hernandez, “Climate change drivers of mosquito-borne disease outbreaks,” Nat. Clim. Chang., vol. 14, pp. 155–165, 2024, doi: 10.1038/s41558-024-01822-9.

[22] F. Marinho, “Climate variability and dengue transmission in Brazil,” Environ. Res., vol. 207, p. 112198, 2022.

[23] R. Torres and M. Velasquez, “Machine learning approaches for tropical disease surveillance,” IEEE Trans. Comput. Soc. Syst., vol. 7, no. 6, pp. 1390–1402, 2020, doi: 10.1109/TCSS.2020.2999911.

[24] Y. Liu, X. Zhang, and H. Chen, “Unsupervised anomaly detection for climate-related infectious disease surveillance using machine learning,” Environmental Modelling & Software, vol. 158, p. 105597, 2023, doi: 10.1016/j.envsoft.2022.105597.

[25] J. Guo and X. Liu, “Using DBSCAN for outbreak anomaly detection in public health surveillance,” IEEE Access, vol. 8, pp. 73422–73433, 2020, doi: 10.1109/ACCESS.2020.2988773.

[26] M. Faisal and T. K. A. Rahman, “Optimally Enhancement Rural Development Support Using Hybrid Multy Object Optimization (MOO) and Clustering Methodologies: A Case South Sulawesi - Indonesia,” International Journal of Sustainable Development and Planning, vol. 18, no. 6, pp. 1659–1669, 2023, doi: 10.18280/ijsdp.180602.

[27] M. Faisal et al., “A Hybrid MOO-MCDM-PSOM Approach for Mapping Urban Carbon Trade Policies,” 2025. doi: 10.20944/preprints202502.0981.v1.

[28] M. G. Teixeira, “Spatiotemporal dynamics of dengue epidemics in Brazil,” PLoS Negl. Trop. Dis., vol. 15, no. 6, p. e0009495, 2021.

[29] W. Hu, “Spatiotemporal dengue outbreak detection using environmental risk indicators,” PLoS Negl. Trop. Dis., vol. 14, no. 3, p. e0008084, 2020, doi: 10.1371/journal.pntd.0008084.

[30] B. Cazelles, M. Chavez, and S. Hales, “Nonlinear climate–disease interactions and early outbreak signals in dengue transmission,” Epidemics, vol. 46, p. 100738, 2024, doi: 10.1016/j.epidem.2024.100738.

[31] N. Thao and P. Hung, “Climate–disease systems with non-linear outbreak dynamics,” Environmental Modelling, vol. 265, p. 105851, 2024, doi: 10.1016/j.envmodel.2024.105851.

[32] Y. Chen and T. Wang, “Unsupervised anomaly detection in epidemiological surveillance systems,” J. Biomed. Inform., vol. 142, p. 104333, 2023, doi: 10.1016/j.jbi.2023.104333.

[33] M. Suarez, “Mosquito ecology and rainfall cycles in dengue-endemic regions,” Infect. Dis. Model., vol. 6, pp. 330–345, 2021, doi: 10.1016/j.idm.2021.01.003.

[34] A. Eskildsen and T. Roslin, “Temperature-dependent shifts in mosquito distribution,” Glob. Chang. Biol., vol. 28, no. 3, pp. 945–957, 2022, doi: 10.1111/gcb.15988.

[35] C. Moreno and L. Johnson, “Global surveillance of infectious diseases using climate indicators,” Lancet Planetary Health, vol. 8, pp. 112–121, 2024, doi: 10.1016/S2542-5196(24)00112-4.

[36] C. Ramos and P. Torres, “Multivariate anomaly detection in climate epidemiology datasets,” Comput. Biol. Med., vol. 159, p. 106123, 2023, doi: 10.1016/j.compbiomed.2023.106123.

[37] A. M. Da Silva, “Rainfall anomalies and dengue fever outbreaks in tropical regions,” Clim. Risk Manag., vol. 33, p. 100349, 2021, doi: 10.1016/j.crm.2021.100349.