A Machine Learning Approach to Analyse the Symptoms of Covid-19 for the Initial Diagnosis of a Patient

Abstract

The recent outbreak of the respiratory ailment COVID-19 caused by novel corona virus SARS- Cov2 is a severe and urgent global concern. In the absence of vaccine, and also treatment of COVID- 19 WHO (World Health Organization) had informed that Social distancing is the only way to avoid this pandemic and also made clear that Prevention is better than Cure. The main containment strategy is to reduce the contagion by the isolation of affected individuals. Earlier stage this pandemic was declared as a sort of Pneumonia where an individual gets affected by cold, fever and headache. Later, some new symptoms are seen in affected people like sore throat, breathing problems, and sometimes constipation. To make rapid decisions on treatment, and isolation needs, it would be useful to determine which symptoms presented by suspected infection cases are the best predictors of a positive diagnosis. This can be done by analyzing patient's symptoms and its outcome. Here, we developed a model that employed supervised machine learning algorithms to identify the certain features predicting COVID-19 disease diagnosis with high accuracy. Features examined includes details of the concerned individual, e.g., age, gender, observation of fever, breathing difficulty, and clinical details such as the severity of cough and incidence of lung infection and congestion. We had implemented some Machine Learning techniques with algorithms and found out the highest accuracy more than (50 %) of individual patient for all age groups. The following data is collected from COVID-19 positive patients, online survey and social survey done at testing centres. After that we had applied various methods as Data Preprocessing, Model Validation and Statistical analysis, etc. The probability and accuracy of a patient is shown in using various methods of Machine learning algorithm for a better understanding.