Indonesian Covid-19 Future Forecasting Based on Machine Learning Approach

Wahyuni, S. N.; Sediyono, E.; Sembiring, I.

Since the Coronavirus disease (COVID-19) outbreak emerged in China, several studies attempt to predict COVID-19 as an effort to reduce the spread of the virus. Indonesia is one of the countries with a Covid-19 death rate is the highest in Southeast Asia. These need early preventive measures, One of them is by forecasting the number of the COVID-19 cases.This is very important to be taken to determine the number of maximum cases as a prepare health facilities effort in Indonesia and preparations for reopening economic activities in Indonesia. The focuses of this study are forecasting future COVID-19 in Indonesia using the Multiple Linear Regression model. The data set were taken from the website: https://data.humdata.org/dataset/indonesia-Covid-19-cases-recoveries-And-deaths-perprovince. The tools using Python. The result of this study is the MLR model has high accuracy in the prediction model. This is shown by several results, among others, first the P-value smaller than α (0.005). Second, the F value is higher than the F table, where the F value is 901.665 and the F table is 3,170234. Third, the t-value of affected and recovered cases is higher than the t table. Where t-value is both of them are 4003 and 15529, meanwhile t table is 1,969. The last R2 value is 0.9999 and the Mean standard Error (MSE) is 0.0048 The predicted results shown have an average difference in results of 0.037%. © 2021 IEEE.