A study and identification of COVID-19 viruses using N-grams with Naïve Bayes, K-Nearest Neighbors, Artificial Neural Networks, Decision tree and Support Vector Machine

Abstract

Coronavirus disease 2019 or COVID-19 is a global health crisis caused by a virus officially named as severe acute respiratory syndrome coronavirus 2 and well known with the acronym (SARS-CoV-2). This very contagious illness has severely impacted people and business all over the world and scientists are trying so far to discover all useful information about it, including its potential origin(s) and inter-host(s). This study is a part of this scientific inquiry and it aims to identify precisely the origin(s) of a large set of genomes of SARS-COV-2 collected from different geographic locations in all over the world. This research is performed through the combination of five powerful techniques of machine learning (Naïve Bayes, K-Nearest Neighbors, Artificial Neural Networks, Decision tree and Support Vector Machine) and a widely known tool of language modeling (N-grams). The experimental results have shown that the majority of the aforementioned techniques gave the same global results concerning the origin(s) and inter-host(s) of SARS-COV-2. These results demonstrated that this virus has one zoonotic source which is Pangolin.