COVID-19 Case Prediction in Indonesia Using Ensemble Machine Learning Techniques with Vaccinations

The COVID-19 pandemic has significantly impacted Indonesia, with recurring waves of infections challenging the country's healthcare and policy responses. Predicting the trajectory of COVID-19 cases is essential for effective public health planning. This study focuses on utilizing ensemble machine learning techniques, including Random Forest, Gradient Boosting, AdaBoost, Bagging, and Voting, to forecast COVID-19 cases in Indonesia. By integrating vaccination data—such as coverage rates and vaccine types—into these models, we aim to enhance the accuracy of predictions. Our evaluation using metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Relative Absolute Error (RAE), and p-value shows that Random Forest achieved the lowest MAE of 68.17 and RMSE of 294.21, making it the most effective model. Gradient Boosting also performed well with an MAE of 254.26 and RMSE of 493.72. However, AdaBoost performed poorly, with the highest MAE of 1782.67 and RMSE of 2055.50. The results indicate that ensemble models, particularly Random Forest and Bagging, provide more reliable predictions with low error rates. These findings contribute to a better understanding of the effects of vaccination campaigns on virus transmission and healthcare strain, offering insights for data-driven strategies in managing COVID-19 in Indonesia.