Using different machine learning models to classify patients into mild and severe cases of COVID‐19 based on multivariate blood testing

• Published in: Journal of medical virology Vol. 94; no. 1; pp. 357 - 365
• Main Authors: Zhang, Rui‐kun, Xiao, Qi, Zhu, Sheng‐lang, Lin, Hai‐yan, Tang, Ming
• Format: Journal Article
• Language: English
• Published: London Wiley Subscription Services, Inc 01.01.2022; John Wiley and Sons Inc
• Subjects: artificial intelligence; Artificial neural networks; Bayesian analysis; biostatistics & bioinformatics; Blood; Blood tests; Classification; coronavirus; COVID-19; Indicators; infection; Learning algorithms; Machine learning; Neural networks; Regression models; Respiratory diseases; Support vector machines; Virology; virus classification
• ISSN: 0146-6615; 1096-9071; 1096-9071
• DOI: 10.1002/jmv.27352

COVID‐19 is a serious respiratory disease. The ever‐increasing number of cases is causing heavier loads on the health service system. Using 38 blood test indicators on the first day of admission for the 422 patients diagnosed with COVID‐19 (from January 2020 to June 2021) to construct different machine learning (ML) models to classify patients into either mild or severe cases of COVID‐19. All models show good performance in the classification between COVID‐19 patients into mild and severe disease. The area under the curve (AUC) of the random forest model is 0.89, the AUC of the naive Bayes model is 0.90, the AUC of the support vector machine model is 0.86, and the AUC of the KNN model is 0.78, the AUC of the Logistic regression model is 0.84, and the AUC of the artificial neural network model is 0.87, among which the naive Bayes model has the best performance. Different ML models can classify patients into mild and severe cases based on 38 blood test indicators taken on the first day of admission for patients diagnosed with COVID‐19. Highlights 1.Different machine learning models are used to classify mild and severe cases of COVID‐19 patients. 2.38 blood sample indicators are used for the machine learning models. 3.Five most significant indicators are identified.