Prediksi Jumlah Pasien Sembuh COVID-19 Menggunakan Jaringan Syaraf Tiruan

Abstract

Diseases experienced by humans can be sourced from various things such as bacteria, viruses, fungi and others. Serious diseases that have been experienced in the world include MERS, SARS, Ebola, HIV/AIDS and many more. In 2019, the World Health Organization (WHO) received a report about a group of patients with pneumonia of unknown cause from the same city, namely Wuhan, the capital of Central China's Hubei Province, named Corona Virus Disease-19 (COVID-19). This COVID-19 has made all affected countries become overprotective of their communities and regions. The number of people who have tested positive for the virus, recovered and even died is increasing every day. One of the public's questions regarding the number of recovered patients is a question that is waiting for an answer because the addition of cured patients is good news for all levels of society in the world. The problem in this study is how to predict the number of recovered patients. This question can be answered with predictions through existing methods, so the purpose of this study is to utilize a Backpropagation Artificial Neural Network which has high accuracy to predict the number of recovered COVID-19 patients in each country with the learning rate, hidden layer, neurons in hidden layers and epochs. The input values for the Backpropagation Neural Network method are the number of positive patients, the number of patients dying, the number of positive cases per day, and the number of cases dying per day. The architecture used is [4-4-1] which means it consists of 4 input values and 4 neurons in the hidden layer and will produce 1 output value. The amount of data used is 171 countries from January 22, 2020 to June 21, 2020 with a total data of 25,992 and produces an average value of 90.9% accuracy.