The Application of Machine Learning in the Detection of Covid-19 in Blood Tests

The disease Covid-19, which is caused by the SARS-CoV-2 virus entering the body, is spreading at a remarkable speed. This virus is so dangerous that it has killed many people in the world since its emergence. This shows that early detection of people with Covid-19 disease is crucial to control the spread of this dangerous virus. Currently, two common diagnostic methods for this disease are CT scan imaging of the lung and RT-PCR molecular test. Regarding the weaknesses of the diagnostic method of molecular testing, we can mention the high cost of conducting the test, dependence on imported kits, and the long time it takes to receive the test results. Although this method has higher diagnostic accuracy than lung CT scan, in this study, blood tests of 199 patients were used. Three machine learning models such as adaptive fuzzy neural network, support vector machine and neural networks were evaluated. The results show that our proposed method can detect people with Covid-19 with an accuracy of 96% and the F1-score method with a percentage of 87%.