Estimation and Prediction of the Omicron COVID-19 Data in Indonesia

The prediction of epidemic spread is an essential aspect of public health planning and resource allocation. This paper discusses an improved prediction method by combining the Extended Kalman Filter (EKF) as an optimal estimation technique with the Moving Average (MA) as a simple time-series prediction technique. The integration of these two methods results in the Modified Extended Kalman Filter (mEKF), which enables prediction even in the absence of actual observation data. This study applies mEKF to the SEIR model using COVID-19 Omicron case data in Indonesia. The results show that mEKF produces prediction values closer to real data compared to MA, which relies heavily on historical patterns. Therefore, mEKF provides more stable and accurate prediction performance, making it suitable for short-term epidemiological forecasting.