Temperature is the dominant predictor of dengue risk in Thailand, and the thresholds the authors recover correspond closely to known biological constraints on vector competence.
Abstract
Background Dengue fever remains a critical public health challenge in Thailand, with transmission dynamics driven by complex interactions between environmental and socioeconomic factors. Understanding these predictive factors is essential for developing robust forecasting systems. Methods We developed a machine learning framework to classify spatiotemporal dengue risk and identify key predictive factors across Thailand. We analyzed 20 years of monthly dengue hemorrhagic fever surveillance data (2003–2022) from 77 provinces, integrating 54 environmental, climatic, and socioeconomic features. We benchmarked four candidate classifiers — logistic regression, support vector machines, random forests, and eXtreme Gradient Boosting (XGBoost) — and selected XGBoost on the basis of performance across six metrics. SHapley Additive exPlanations (SHAP) were used to interpret feature contributions. The dataset was stratified into training (2003–2016) and testing periods, with the latter subdivided into pre-COVID-19 (2017–2019), COVID-19 (2020–2021), and post-COVID-19 (2022) phases. Results The XGBoost model achieved an AUC of 0.80 in pre-pandemic testing and 0.74 across the combined during- and post-pandemic period. Temperature dominated the feature-importance ranking, comprising seven of the top ten features, with non-linear thresholds near 21°C for 1-month lagged minimum temperature and near 32°C for 3-month lagged maximum temperature — values that align with established biological constraints on Aedes aegypti–mediated transmission. Precipitation features contributed minimally to model predictions, while a higher Gross Provincial Product was associated with increased dengue risk, consistent with predominantly urban transmission patterns. Model performance deteriorated significantly during the COVID-19 pandemic (AUC = 0.62 in 2021), with systematic overprediction indicating that non-environmental factors operating outside the model dominated dengue dynamics during this period. Conclusions Temperature is the dominant predictor of dengue risk in Thailand, and the thresholds we recover correspond closely to known biological constraints on vector competence. Environmentally driven prediction is reliable under stationary conditions but degrades substantially during periods of major societal disruption, underscoring the need to integrate behavioral and surveillance-coverage indicators alongside environmental predictors when applying such models in real time.
The feasibility of predicting dengue outbreaks using weekly multivariate time-series data from San Juan, Puerto Rico and Iquitos, Peru is examined, demonstrating that advanced machine learning models can serve as reliable early warning systems for vector-borne diseases.
Pratik S. Machchar, Purvi N. Ramanuj, R. Patel et al.· International Journal of Inf...· 0 citations
Background Dengue fever poses a pervasive, yet escalating public health burden in Mexico and abroad. Methods We conducted a 41-year spatiotemporal analysis of dengue fever across Mexico (1985–2025), integrating monthly case surveillance with climate, land cover, vegetation, and novel disaster severity covariates derived from the Emergency Events Database. Four supervised regression models were trained on 1990–2021 data, with models evaluated on a 2022–2023 temporal holdout and against observed 2024–2025 surveillance totals. Five supplementary hazard analyses examined temporal correlation, disaster type breakdown, spatial co-occurrence, pre/post event trajectories, and sensitivity to scoring weight assumptions. Results Mann–Kendall trend analysis identified statistically significant increasing dengue incidence in 15 of 32 states (9 inland), evidencing geographic expansion over four decades. Z-score analysis confirmed 2024 as a profound anomaly across both endemic and emerging states. Hazard features were significantly associated with national monthly dengue counts at lags 0–2 months across the full 1985–2025 series. A + 613% case increase following the June 2024 tropical storm. Sensitivity analyses confirmed the project’s developed ‘Severity Score’ performed comparably to four theoretically motivated differential weighting schemes. Conclusion This study provides the first structured integration of disaster severity covariates in a dengue forecasting framework for Mexico. Geographic expansion into inland states, the 41-year trend analysis, and the hazard adjustment results collectively support a differentiated public health response that targets both endemic coastal states and newly affected inland regions. The analytical framework is directly transferable to emerging dengue risk contexts in the United States and Central America, geographic neighbors also experiencing increased dengue virus transmission.
Huixuan Li, Christopher Lee, Sean Sweeney et al.· Frontiers in Public Health· 0 citations
Dengue is an arboviral disease of high public health relevance, characterized by pronounced temporal variability, nonlinearity, and recurrent outbreaks, which pose challenges to epidemiological surveillance and decision-making. This study evaluated the performance of machine learning methods for short-term forecasting of the weekly dengue morbidity rate in the 27 Brazilian capital cities, comprising the 26 state capitals and Brasília, Federal District, with horizons up to 4 weeks. Epidemiological, climatic, and socioeconomic data were compiled for these capital cities and used to compare a Gated Recurrent Unit (GRU) neural network, formulated as a Multi-Input Multi-Output (MIMO) model, and a Gradient Boosting model (CatBoost), implemented using a Direct forecasting strategy with horizon-specific models. Validation was conducted using a walk-forward approach, with evaluation based on absolute error metrics and the coefficient of determination. The results indicated that the GRU architecture presented recurring limitations, including underfitting, temporal lag, and low capacity to anticipate epidemic peaks. In contrast, the CatBoost model demonstrated greater robustness and better adaptation to the variability of epidemiological time series, showing superior performance in most of the analyzed capitals. The findings reinforce that greater architectural complexity does not necessarily imply better operational performance and highlight the potential of ensemble-based methods for short-term epidemiological surveillance applications. These findings contribute to dengue forecasting by showing that, under a common validation framework, ensemble-based strategies may provide greater operational robustness than recurrent MIMO architectures for short-term prediction in heterogeneous epidemiological settings.
D. C. da Cunha e Silva, L. M. Nery, Nícholas de Paula Nicomedes et al.· International journal of bio...· 0 citations
Meteorological factors influence dengue transmission through nonlinear and lagged effects. This study assessed these associations in Hainan Province and developed an early risk prediction index.
Weekly dengue case and meteorological data from Hainan Province (2018–2019 and 2023–2024) were analyzed using distributed lag non-linear models (DLNMs). An early risk prediction index(ERPI) was constructed from DLNM-derived cumulative meteorological risk signals and imported cases. Predictive performance was evaluated using leave-one-year-out cross-validation and assessed by the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).
Among 927 dengue cases reported during the four modelled years, 727 (78.43%) were locally acquired. Weekly mean daily maximum temperature (Tmax), precipitation (Pmean), wind speed (WSmean), and diurnal temperature range (DTRmean) showed nonlinear, lag-dependent associations with dengue risk, with distinct temporal patterns. Higher Tmax showed longer-lag effects, peaking at 33.73 °C at lag 8 weeks (
RR
= 2.45, 95% CI: 1.70–3.63). Its cumulative effect was strongest over lags 5–8 weeks, peaking at approximately 32 °C. Pmean showed a nonlinear association, peaking at 4 mm/day at a lag of 4 weeks (
RR
= 1.92, 95% CI: 1.59–2.31), with significant cumulative effect across both the 1–4- and 5–8-week windows. Low WSmean and moderate DTRmean showed their strongest cumulative associations over lags 1–4 weeks, peaking at 2.0 m/s and 6.1 °C, respectively. These harmful associations attenuated over lags 5–8 weeks, whereas higher WSmean and DTRmean were generally protective. The ERPI predicted local dengue occurrence within the subsequent 1–4-week windows, with AUCs of 0.762–0.843. Across these windows, sensitivities were 0.768–0.797, specificities were 0.644–0.708, PPVs were 0.468–0.624, and NPVs were 0.822–0.911.
In Hainan, high temperature was primarily associated with medium-to-long-term dengue risk, whereas low-to-moderate diurnal temperature range increased short-term risk. Precipitation was associated with risk across both short and medium-to-long lags, while cumulative wind-speed associations were strongest over shorter lags and attenuated thereafter. The DLNM-based early risk prediction index may support meteorology-informed early warning and dengue control in tropical settings, although requiring external validation is needed before operational implementation.
Unknown authors· Frontiers in Public Health· 0 citations
The experimental results show that climate-aware LSTM models are good at predicting dengue cases and can help build strong early warning systems in areas where cases are dependent on climate change.
Amiben Mehta, Kajal Patel· ITEGAM- Journal of Engineeri...· 0 citations
Findings indicate that SARIMAX is more suitable for forecasting dengue incidence characterized by strong seasonal patterns and relatively limited observations, than LSTM.
George Elmar, Asriyanik, Winda Apriandari· Kontribusia (Research Dissem...· 0 citations