A Predictive Model for Data-Driven Insights into MonkeyPox Virus Outbreaks and Case Trends in Africa

Abstract

Human monkeypox (MP) is a zoonotic disease caused by the monkeypox virus (MPV). It is considered one of the most significant orthopoxvirus infections after smallpox. MPV originated from the Congo Basin and West African clades, MP virus has spread globally, with recent outbreaks highlighting the need for predictive models in understanding its transmission. The aim of the study is to deploy machine learning models for monkeypox outbreak prediction across Africa. In this work, the most recent monkeypox dataset was evaluated and the significant instances were visualised. Feature extraction techniques like Recursive Feature Elimination (RFE), and Least Absolute Shrinkage and Selection Operator (LASSO) were deployed. These methods identify key predictors of monkeypox cases, including total cases, new cases smoothed, and total deaths, to improve model accuracy and interpretability. The various machine learning models employed are Random Forest, Decision Tree, Support Vector Machine, XGBoost, LightGBM, and CatBoost to evaluates their effectiveness in outbreak prediction. The results indicate that Random Forest and XGBoost performed best, achieving accuracy scores of 0.9696 and 0.8949, respectively, with R² values near 1.0 and low RMSE values of 1.29 and 5.51 for Random Forest and XGBoost respectively. The study showed that Random Forest and XGBoost are reliable tools for understanding monkeypox virus transmission dynamics across African countries. These models provide valuable insights for public health interventions and help to identify trends and factors influencing outbreaks.