Improved Smart Forecasting Model to Combat Coronavirus using Machine Learning

Coronavirus (COVID) is one of the world's most devastating diseases, affecting the lives of millions of people around the globe. Accurate and timely detection of the COVID virus is critical for human survival. The standard medical history of diagnosing COVID disease has been deemed unreliable in several ways. Noninvasive methods such as Machine Learning (ML) are efficient and reliable for classifying healthy persons and people with COVID disease. In this paper, we improved a Smart Forecasting Model using Machine Learning (SFML) for COVID prediction (positive, negative) by using COVID data. We applied, two algorithms of feature selection (correlation and gain ratio), two supervised ML algorithms (Random forest and support vector machine) and the technique of cross-validation was applied for evaluating the SFML, such as specificity, accuracy, F-measure, sensitivity, and running time. The proposed SFML may easily distinguish COVID infected individuals from healthy individuals. The proposed SFML has been tested on a full set of COVID features as well as a smaller set of features. The results reveal that reducing COVID features has an impact on SMFL performance measures.