Covid-19 Survival Prediction and Diabetes Mellitus relevance using Cox Regression

Abstract

COVID-19 and related viruses have spread fresh dangers to our society from all over the world There is a strong desire to put in place precautionary measures to help prevent outbreaks. The COVID-19 pandemic's impact has generated a flurry of research aimed at better understanding, tracking, and managing the disease. Diabetes mellitus is one of the most prevalent neurological illnesses in COVID-19 patients, and it is linked to poor disease outcomes. In the field of medical diagnosis, machine learning is becoming more common. It can be attributed largely to advancements in disease classification and identification systems, which can provide evidence that assists medical experts in the early detection of deadly diseases, resulting in a dramatic rise in patient survival rates. In this paper, we would like to introduce a method of predicting the probability of survival of an individual infected with COVID-19, which is troubling and widely distributed in the current case scenario, using recent algorithmic improvements that were created to predict death and recovery rates. Our model aims to transfer learning, model integration, and classify a person's probability of survival based on a number of variables and parameters, including death and recovery rates using cox regression and baseline hazard concept for died as input and recovered as input respectively.