Classification of Covid-19 RT-PCR Test Results Using Auto-encoder And Random Forest

Abstract

Corona Virus Disease (COVID-19) is a new type of virus that emerged at the end of 2019. COVID-19 has become a pandemic due to the increase in the number of cases taking place very quickly and has spread to all corners of the world. The World Health Organization (WHO) recommends the use of the Reverse Transcription-Polymerase Chain Reaction (RT-PCR) method as a way to test the diagnosis of COVID-19 infection. This study builds a classification system for the COVID-19 RT-PCR test results by applying the Auto-encoder algorithm and the Random Forest classification. The dataset used is the result of the RT-PCR test from one of the hospitals in Brazil. The method used is the Auto-encoder to process the dataset features first and the Random Forest algorithm to classify the RT-PCR test results that have positive and negative labels. From this process, it can be seen that the Auto-encoder model can process datasets well and the classification carried out using Random Forest can classify with an accuracy of 87.2%.