Estimating SIR Model Parameters for Predicting Covid-19 Cases in Bandung City
Kahfi Mohamad Bintang, Norma Binti Alias, Wandi Yusuf Kurniawan, Ni Kadek Winda Patrianingsih, Gde Palguna Reganata, Putu Harry Gunawan · 2024
The COVID-19 pandemic has profoundly impacted global public health, with Indonesia among the most severely af-fected. In response, the Bandung City Government implemented various control measures, including the Mayor's Circular on COVID-19 prevention and enforcing restrictions on community activities, known as PPKM. This study applies the Susceptible-Infected-Recovered (SIR) model to analyze the progression of COVID-19 in Bandung City, specifically during the first and second waves. The SIR model was calibrated using historical case data and incorporated the finite probability of recovered individuals becoming susceptible again, which is critical for generating oscillatory solutions in epidemic modeling. The study focused on determining the best transmission rate for both waves. The first wave, occurring between July 4th and October 22nd, 2021, exhibited an optimal transmission rate of 0.38 (MAE: 0.0424). In contrast, the second wave, spanning February 1st to May 15th, 2022, required a higher transmission rate of 0.54 (MAE: 0.0142), reflecting the increased transmissibility of the Omicron variant. Meanwhile, the recovery rate was calculated based on an average incubation period of 14 days, equivalent to 1/14, approximately 0.0714. The results demonstrate that the SIR model successfully captures the epidemic dynamics in the region, providing insights into the effectiveness of governmental policies and restrictions in controlling the virus. Future studies may explore the inclusion of additional parameters, such as vaccination coverage and mobility patterns, which could enhance the model's predictive accuracy. Moreover, hybrid modeling approaches integrating SIR with machine learning may offer more adaptive solutions to rapidly changing epidemic conditions.