Early Detection of COVID-19 in Patients with Comorbidities using a Novel Deep Learning Model

- COVID-19 has become among the most severe and enduring illnesses of recent times because of its widespread
distribution. When the sickness has been more broadly dispersed, it is difficult to tell who was actually impacted. Over sixty percent of impacted people claim to have a dry cough. Sneezing and other respiratory noises have been used to create diagnostic models in numerous recent research. Applications for deep learning (DL) in healthcare seem revolutionary. DL makes use of neural networks to boost processing power and produce reliable results. With this cutting-edge medical technology, doctors can accurately analyze any ailment, allowing them to treat it more effectively and, as a result, make better medical judgements. This research proposed a novel DL algorithm, i.e., Bifold Long Short-Term Memory, for detecting COVID-19 infection (BFLLCOV-19) in individuals who may have the possibility of infection with or without comorbidities. This research work acquires datasets received through custom-designed online Google forms and data received from individuals. The COVID-19 pandemic outbreak is, without a doubt, the worst disaster of the twenty-first century and likely the most important worldwide crisis that hit great nations economically. The virus's propensity to spread quickly has forced the global populace to maintain tight protection measures to prevent self and slow down the disease's