Fig. 3
From: Reconstructing the cytokine view for the multi-view prediction of COVID-19 mortality
The mortality prediction of COVID-19 patients. This experiment used three feature selection algorithms, i.e., AdaBoost, RFE-SVM, and RF. The classification models were trained using five classifiers, i.e., NB, CART, KNN, RF, and SVM. The data series B, C, and B + C represent the models using complete blood counts only, cytokine levels only, and both datasets, respectively. Each model was trained using the dataset A1-Training and tested on the dataset A1-Testing.