Implementasi Sistem Pembatas Kapasitas Pengunjung berdasarkan Pendeteksian Gejala Suspek COVID-19 menggunakan Metode Random Forest

Rahmat Yusuf Afandi (Fakultas Ilmu Komputer, Universitas Brawijaya)
Rizal Maulana (Fakultas Ilmu Komputer, Universitas Brawijaya)
Eko Setiawan (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 6 No 9 (2022): September 2022

Monitoring the number of visitors in public places that apply PPKM rules is difficult to do manually if in a short time too many visitors come and go. Checking symptoms of suspected COVID-19 cases in public places using only visitors' body temperature parameters is still less accurate considering the symptoms of suspected COVID-19 have several parameters that can be measured. This study developed a system that can calculate visitor capacity automatically and can measure symptoms of suspected COVID-19 in three parameters, namely body temperature, oxygen saturation in the blood, and respiration time. The measurement results of the three parameters will be classified into two classes using the Random Forest method. The system output is displayed by LCD containing text of the sensor measurement value and the classification results as well as the sound produced by buzzer containing an appeal to implement health protocols. The results of oxygen saturation measurement test using the MAX 30100 sensor obtained an average accuracy of 98.98%. Measurement test of body temperature using the MLX 90614 sensor obtained an average accuracy of 99.29%. Measurement test of respiration time using the KY-037 sensor obtained an average accuracy of 97.65%. Measurement test of visitor capacity using the PIR sensor obtains an accuracy of 100%. The test results of fifteen test data against the Random Forest classification achieved an accuracy of 100% with an average computation time of 35.3 s. Monitoring the number of visitors in public places that apply PPKM rules is difficult to do manually if in a short time too many visitors come and go. Checking symptoms of suspected COVID-19 cases in public places using only visitors' body temperature parameters is still less accurate considering the symptoms of suspected COVID-19 have several parameters that can be measured. This study developed a system that can calculate visitor capacity automatically and can measure symptoms of suspected COVID-19 in three parameters, namely body temperature, oxygen saturation in the blood, and respiration time. The measurement results of the three parameters will be classified into two classes using the Random Forest method. The system output is displayed by LCD containing text of the sensor measurement value and the classification results as well as the sound produced by buzzer containing an appeal to implement health protocols. The results of oxygen saturation measurement test using the MAX 30100 sensor obtained an average accuracy of 98.98%. Measurement test of body temperature using the MLX 90614 sensor obtained an average accuracy of 99.29%. Measurement test of respiration time using the KY-037 sensor obtained an average accuracy of 97.65%. Measurement test of visitor capacity using the PIR sensor obtains an accuracy of 100%. The test results of fifteen test data against the Random Forest classification achieved an accuracy of 100% with an average computation time of 35.3 s.