TfrAdmCov: a robust transformer encoder based model with Adam optimizer algorithm for COVID-19 mutation prediction

Abstract

ABSTRACT The development of vaccines and drugs is very important in combating the coronavirus disease 2019 (COVID-19) virus. The effectiveness of these developed vaccines and drugs has decreased as a result of the mutation of the COVID-19 virus. Therefore, it is very important to combat COVID-19 mutations. The majority of studies published in the literature are studies other than COVID-19 mutation prediction. We focused on this gap in this study. This study proposes a robust transformer encoder based model with Adam optimizer algorithm called TfrAdmCov for COVID-19 mutation prediction. Our main motivation is to predict the mutations occurring in the COVID-19 virus using the proposed TfrAdmCov model. The experimental results have shown that the proposed TfrAdmCov model outperforms both baseline models and several state-of-the-art models. The proposed TfrAdmCov model reached accuracy of 99.93%, precision of 100.00%, recall of 97.38%, f1-score of 98.67% and MCC of 98.65% on the COVID-19 testing dataset. Moreover, to evaluate the performance of the proposed TfrAdmCov model, we carried out mutation prediction on the influenza A/H3N2 HA dataset. The results obtained are promising for the development of vaccines and drugs.