Predictive Analysis of Monkeypox Case Trends Through Machine Learning Models: A Feature driven Approach Using Model Interpretability

Monkeypox is an infectious disease caused by the monkeypox virus and has symptoms similar to smallpox, but typically less severe. The world is, at the moment, battling against the epidemic fear of the prevalence of monkeypox outbreaks in various corners of the globe, especially in Africa, and throughout the Western hemisphere. The term "Monkeypox" encapsulates the disease perfectly because it was first detected in Macaque monkeys. This study aims to analyze the average new cases of monkeypox and evaluates the effectiveness of managing these cases using statistical analysis and machine learning techniques. This research employs statistical analysis to assess the feature dependency and association of the features. Additionally, this study provides a correlation heatmap inspecting the relationship of each variable. Moreover, machine learning models such as XGBoost, Random Forest, GRB, LightGBM Regressor, CatBoost Regressor, and KNN were trained with highly aligned features and achieved expected predictions. Furthermore, SHAP (Shapley Additive Explanation) and LIME (Local Interpretable Model-agnostic Explanation) were used to obtain a better visualization of predicted models. It is very much probable that increasing R2-score (92.22) and decreasing MSE (0.171) encapsulate the model’s effective predictions. Moreover, for further research, it is important to collect a dataset with a higher number of features.