A Deep Learning Approach to Outbreak Virus Classification: Utilizing Bidirectional GRU on DNA Sequence of SARS-CoV-2, Zika, Ebola, and MERS

Over the past decades, at least 4 types of virus have caused global outbreaks and pandemics: SARS-CoV-2, Zika, Ebola, and MERS. In order to reduce losses, researcher manage to classify the virus. DNA sequence plays an essential role in defining the specific virus, and the sequence itself must be accurate. Processing virus data on a large scale also requires the integration of computer science to help speed up the analysis and processing of DNA sequence data. In this instance where the model must comprehend the context of each sequence, BiGRUs are used because it has better data understanding ability in order to generate correct predictions. Using Bi-GRU and N-Mers Frequency, outbreak virus classification based on DNA Sequence has been made. The model successfully classifying between outbreak virus and also resulting good performance. It has 100% of Accuracy, 100% of F1-Score, 100% of Recall, and 100% of Precision on all data (training, validation, and testing data). This model later will be useful in determining the most suitable treatments for a detected virus-caused disease while also providing new understanding for bioinformatics research.