A Novel Customized Sequential Deep Learning Model (CSDLM) For Covid-19 Risk Prediction

Abstract

One of the most devastating pandemics in human history was caused by the coronavirus 2019 (COVID-19). In order to initiate early intervention, we sought to forecast a bad prognosis among severe patients. It is more likely that the virus may mutate, leading to the formation of highly replicating pathogenic forms, as the number of probable COVID-19 cases rises everyday. A Deep Learning model predicting high risk for COVID-19 from a patient's present symptoms, status, and medical history is the primary goal of this study. We used a Customized Sequential Deep Learning represent (CSDLM) to represent the highly risky COVID-19 situations. A thick layer in this model indicates that every neuron is related to every neuron in surrounding levels, and each layer is closely connected to the one before it. A wide range of measures, including accuracy, precision, recall, and F1_score, were used to thoroughly assess the top performance of the suggested model. This comprehensive assessment approach shed light on how well and how resilient the model predicted COVID-19 high-risk scenarios.