PREDIKSI KESEMBUHAN PASIEN COVID-19 DI INDONESIA MELALUI TERAPI MENGGUNAKAN METODE NAÏVE BAYES

Authors

  • Okky Putra Barus Universitas Pelita Harapan
  • Anton Tehja Universitas Pelita Harapan

Abstract

This study aims to predict the recovery of COVID-19 patients in Indonesia by using Data Mining calculations. The method used to predict the recovery of COVID-19 patients is the Naïve Bayes method. The collection of datasets through trusted sources, the NIHR Innovation Observatory and datasets on an international/global scale, totaling 367 pieces of raw data that have not been filtered. After conducting the data feasibility test, the remaining 286 pieces of data will be divided into 70% of training data of 200 pieces of data and 30% of testing data of 86 pieces of data. Based on the test results, the use of the Naïve Bayes method in predicting the recovery of COVID-19 patients obtained an Accuracy of 96.51%, a Success Precision (Yes) of 100% and a Failure (No) of 95.71%, and a Success Sensitivity (Yes) of 84.21% and Failed (No) by 100%. Therefore, it is concluded that calculations using the Naïve Bayes method in this study will produce an accuracy rate of COVID-19 recovery of 96.51%, which means that the results of the predictions’ calculation of success and failure in a therapy given to patients can be accounted for as data reference in a more detailed subsequent research..

Published

2021-07-26