Fig. 6From: Visual transformer and deep CNN prediction of high-risk COVID-19 infected patients using fusion of CT images and clinical data Visual representations from the t-SNE approach using A: the main dataset including all 67 clinical labels (with no dimension reduction), B: 13 selected clinical labels from SelectKBest, C: 30 selected clinical labels from ExtraTree classifier, and D: 25 extracted features from PCA. blue: high-risk (class 0), orange: low-risk (class 1)Back to article page