A literature review of recurrent neural network approaches to malaria outbreak prediction in Sub-Saharan Africa

Explainable Graph Neural Networks for Multimodal Malaria Risk Prediction Using Malaria Atlas Project (MAP)-Inspired Synthetic Data and Climate Indicators in Sub-Saharan Africa

Malaria remains one of the most significant infectious disease burdens in Sub-Saharan Africa, placing sustained pressure on public health systems and malaria-control programmes. Accurate and interpretable identification of high-risk districts is therefore essential for early warning, resource allocation, and targeted intervention planning. This study proposes an explainable multimodal artificial intelligence framework that integrates epidemiological, climatic, geospatial, and demographic information through a modality-specific encoder-fusion architecture. The framework is further extended with a spatial Graph Neural Network component designed to capture inter-district dependencies and potential spatial spillover effects. Using a synthetically generated district-level dataset (n = 1,200), calibrated to reflect the statistical properties of variables commonly used in Malaria Atlas Project (MAP), NASA POWER, CHIRPS, WorldPop, and Sentinel-2 products across five malaria-endemic countries (the Democratic Republic of the Congo, Uganda, Nigeria, Tanzania, and Kenya), the proposed framework achieved a validation R² of 0.7285 for Plasmodium falciparum parasite rate regression. For binary high-risk classification, the framework attained an area under the curve of 0.9740 and an F1-score of 0.9630 on the held-out test dataset. SHapley Additive exPlanations were used to examine the contribution of individual predictors at global and local levels. Incidence rate (67.2%), rainfall (10.2%), and temperature (6.2%) emerged as the leading drivers of predicted risk, while intervention coverage showed a protective contribution. The results should be read as a synthetic proof-of-concept rather than as an operational malaria map. Even so, they show how multimodal learning, graph-based spatial modeling, and explainable AI can be combined into a transparent surveillance pipeline for future validation with real-world malaria data in resource-constrained settings.

Comparative evaluation of machine learning strategies for short-term dengue forecasting in Brazilian capital municipalities

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.

Data-Driven Geospatial Modeling and Forecasting of Malaria Burden to Support Control and Elimination in Africa

Case-Based Reasoning Model for Predicting the Malaria Cases

Malaria remains a major public health issue in Ivory Coast, where the need for accurate and interpretable predictive models is critical for effective disease control. While most existing approaches prioritize predictive accuracy over interpretability, this study addresses the need for explainable models suitable for deployment in resource-limited public health settings. The study used a dataset covering multiple regions of Cˆote d’Ivoire over the period 2019–2023 and including climatic variables such as temperature, humidity, and precipitation. Seven conventional machine learning models (CatBoost, XGBR, RFR, DTR, SVM, KNN, and Linear Regression) were compared with three proposed Case-Based Reasoning variants (CBR 0, CBR 1, and CBR 2), which differ in their similarity-weighting strategies and correction constant. The results show that CBR 2 achieved the best predictive performance, with RMSE = 0.081, MAE = 0.057, and R2 = 72.40%, followed by the Random Forest Regressor. A Wilcoxon signed-rank test confirmed a statistical significant difference of this permofance (W = 18029, p = 1.02 × 10−20). Beyond predictive accuracy, qualitative criteria including explain-ability, transparency, and adaptability were evaluated, further highlighting the superiority of CBR 2 over conventional black-box models. These findings highlight the potential of Case-Based Reasoning for epidemiological forecasting and decision support in malaria control.

Comparative analysis of single and hybrid time series models for forecasting of deaths from cancer in Somalia

Cancer mortality is rising rapidly across the African continent, yet Somalia lacks localized national forecasting evidence due to decades of instability and the absence of a formal registry. This study aims to develop and compare high-precision time-series forecasting models to project cancer mortality trends in Somalia through 2030, providing a critical statistical baseline for proactive healthcare resource allocation. Annual cancer mortality data (1980–2021) were retrieved from the Global Burden of Disease (GBD) database. A rigorous comparative analysis of 12 representative forecasting configurations was conducted, comprising six single models (including ARIMA, ETS, TBATS, and the Theta method) and six optimized hybrid ensembles integrated with Artificial Intelligence (AI) proxies like Extreme Learning Machines (ELM) and Neural Network Autoregression (NNAR). To stabilize non-stationary trends and mitigate overfitting in a limited sample context (n = 42), second-order differencing (d = 2) and a 10-year rolling-origin cross-validation (tsCV) were implemented. Model performance was benchmarked using RMSE, MAPE, Theil’s U, Willmott’s Index of Agreement (WI), and Skill Scores. All evaluated configurations achieved Theil’s U statistics below 1.0, indicating superior performance relative to the naive benchmark. The parsimonious ARIMA (2,2,1) model emerged as the most accurate configuration for long-term forecasting, achieving the lowest MAPE (1.127%) and RMSE (100.16) in the independent holdout set. The statistical superiority of ARIMA (2,2,1) was confirmed via the Diebold–Mariano test (p < 0.01) against the naive benchmark, outperforming complex hybrid ensembles which showed higher error accumulation over the 8-year horizon. Projections indicate a consistent monotonic upward trend, with annual cancer deaths estimated to reach approximately 9,652 by the year 2030. The findings indicate a significant and ongoing epidemiological transition in Somalia. These forecasts provide a robust statistical baseline to assist the Somali Ministry of Health in strategic planning for oncology infrastructure, diagnostic capacity expansion, and the formal establishment of a National Cancer Registry to transition from reconstructed estimates to actual patient records.

Avian influenza risk mapping in India using machine learning

Highly pathogenic avian influenza (HPAI) remains a recurrent threat to poultry production and One-Health surveillance in India. We developed a national relative spatial risk map for India using curated outbreak records spanning January 2006 to April 2024 (predominantly HPAI H5N1 and H5N8), buffered pseudo-absence sampling, H3 resolution-7 hexagons, and 94 environmental, livestock, land-cover, and anthropogenic predictors. Under 5-fold spatial block cross-validation, eight base classifiers were trained and all performed above chance (AUC > 0.76). The Gaussian-process stacked meta-learner achieved the highest AUC (0.853), but the improvement over the strongest individual base learner, Random Forest (AUC 0.851; Brier 0.155; ECE 0.079), was small and statistically non-significant. Its principal added value was a companion uncertainty layer, the GPR posterior standard deviation, which showed internal consistency with ensemble disagreement across base models (r = 0.74, p < 0.001). Feature attribution ranked human population density, extensive chicken density, June precipitation, and seasonal humidity variables among the predictors most associated with model outputs, with interpretation constrained by passive-surveillance bias and multicollinearity. The resulting relative spatial risk surface, prediction-uncertainty surface, and subdistrict risk-uncertainty classification layers identify eastern, northeastern, coastal, and selected southern regions as priorities for targeted surveillance and prospective validation.