Table 4 Pneumonia subtype discriminability of the machine learning model.

From: Machine learning application for the prediction of SARS-CoV-2 infection using blood tests and chest radiograph

Disease

Positive/total

AUC, % (CI)

Accuracy, % (CI)

Sensitivity, % (CI)

Specificity, % (CI)

N

% (95%-CI)

% (95%-CI)

% (95%-CI)

% (95%-CI)

Bacterial PNA

45/175

77.4 (70.2–84.5)

55.4 (47.7–62.9)

100.0 (92.1–100.0)

40.0 (31.5–49.0)

Viral PNA

49/175

62.9 (54.0–71.9)

56.6 (48.9–64.0)

67.3 (52.5–80.1)

52.4 (43.3–61.3)

Non PNA

62/175

62.5 (54.1–70.9)

61.7 (54.1–68.9)

38.7 (26.6–51.9)

74.3 (65.3–82.1)

  1. AUC area under the curve, PNA pneumonia.
Back to article page