Diagnosis of COVID19 by Using Machine Learning

This study aimed to conduct intensive experiments on predictive machine learning models to develop and compare prognosis to highlight the parameters behind the death due to corona virus (Covid19). Datasets were obtained from Mexican Government. Various machine learning algorithms were used to extract the hidden patterns from the obtained datasets. Feature selection technique was employed to optimize the accuracy of the classification algorithms. Although the dataset was big, the 10-fold Cross Validation technique was used to validate the experimental results. Several metrics were used to measure the accuracy of the algorithms used and then compare between the results of the conducted experiments. Based on the analysis of multiple algorithms, it has been found that J48 algorithm has the best classification performance compared to other machine learning algorithms. More importantly, it has been demonstrated that some parameters have significant contributes to the death of the infected patients. For predicting mortality, oxygen saturation, pneumonia and age are the leading predictors. The obtained results are promising as they give an insight knowledge about the main causes of patient status: recovery and death.