Prediksi Jumlah Kasus COVID-19 Menggunakan Teknik Sliding Wondows dengan Metode BPNN

Abstract

COVID-19 is a disease that attacks the respiratory system caused by the coronavirus. The spread of COVID-19 has made the world restless, including Indonesia. Forecasting the number of COVID-19 cases in Indonesia needs to be done because it is expected to prevent or reduce the number of COVID-19 cases in Indonesia. One of the methods used in this research is to predict the number of COVID-19 cases using Backpropagation Neural Network and Sliding Windows by utilizing daily historical data on the number of COVID-19 cases in Indonesia. The data used in this study is historical data on the number of cases from March 2020 to April 2021. The historical data for the addition of the number of cases is displayed using the Sliding Windows concept approach based on the window sizes used, namely 2, 3, 4, and 5. Windows size reflects the number of days as an input layer variable on the Backpropagation architecture to predict the number of cases the next day. The prediction results obtained with the smallest Sum Squared Error value in network testing is 0.69 as a high predictive accuracy value. The network architecture with the smallest SSE using 3 input layers, 6 hidden layer neurons, and 1 output layer can be knowledge that can help the government in predicting the number of COVID-19 cases in Indonesia in the future.