Stacked ensemble modelling of imported malaria from African countries to China

Imported malaria has been a risk of malaria reemergence in countries that have successfully eliminated malaria. This study developed an ensemble machine learning model to analyze and predict imported malaria cases from African countries to China based on data collected from 2012 to 2018. The stacked ensemble model which combines LASSO, Random forest, and Decision Tree regression, showed a significant improvement in performance, achieving an R 2 range of 0.82 to 0.98, indicating high predictive accuracy, and outperformed individual base models, with a gain of approximately 15% in MAE reduction, highlighting its effectiveness. The findings indicate that the increase in Chinese workers, foreign direct investment (FDI) and existing malaria incidence in African countries associate significantly with the rise in imported malaria cases (P = 0.001). The results underscore the importance of collaboration between China and African nations to address malaria, particularly in formulating targeted public health strategies to mitigate imported cases linked to labor movements. This predictive tool can assist in guiding health interventions and resource allocation in China as intercountry movements continue to grow.