PREDICTION OF COVID-19 SEVERITY IN SARS-COV-2 RNA-POSITIVE PATIENTS BY DIFFERENT ENSEMBLE LEARNING STRATEGIES

Introduction: While the coronavirus only persists marginally for 95% of the infected cases, the remaining 5% are in critical or life-threatening conditions. This study aimed to design an intelligent model that predicts the severity level of the disease by modeling the relationships between the COVID-19 infection severity and the various demographic/clinical features of individuals. Materials and methods: A public dataset of a cross-sectional study including the demographic and symptomatological characteristics of 223 COVID-19 patients was used and randomly partitioned into training (75%) and testing (25%) datasets. During training, the class imbalance problem was solved, and the related factors with the COVID-19 severity were selected using the evolutionary method supported by a genetic algorithm. Neural Network (NN), Support Vector Machine (SVM), QUEST algorithms together with confidence weighted voting, voting, and highest confidence wins strategies (HCWS) were constructed, and the predictive power of models was determined by performance metrics. Results: Based on the performance indicators, among the individual models, the NN model outperformed SVM and QUEST algorithms in the training and testing datasets. However, ensemble approaches gave better predictions as compared to individual models according to all the evaluation metrics. Conclusion: The proposed voting ensemble model outperforms other ensemble and individual machine learning approaches for the severity prediction of COVID-19 disease. The proposed ensemble learning model can be integrated into web or mobile applications to classify the severity of COVID-19 for clinical decision support.

Automated summary

This study aimed to design an intelligent model to predict the severity level of COVID-19 by modeling the relationships between infection severity and demographic/clinical features. Results showed that the proposed voting ensemble model outperformed other individual and ensemble machine learning approaches. The model can be integrated into web or mobile applications for clinical decision support.