Table 3 Detailed results of machine learning-based classification of 5-year major adverse cardiovascular and cerebral events [95% confidence interval].

From: Prognostic value of texture analysis from cardiac magnetic resonance imaging in patients with Takotsubo syndrome: a machine learning based proof-of-principle approach

Machine learning classifier

Sensitivity %

Specificity %

Precision

Recall

F-Measure

AUC from ROC curve analysis

PRC Area

ANN

85.2 [82.9–87.6]

17.4 [15–20]

0.94 [0.9–0.97]

0.85 [0.83–0.88]

0.9 [0.86–0.94]

0.79 [0.74–0.84]

0.94 [0.93–0.96]

J48 (C4.5)

85.1 [82.5–87.6]

17.3 [15–20]

0.94 [0.9–0.98]

0.85 [0.82–0.88]

0.9 [0.85–0.94]

0.51 [0.48–0.53]

0.84 [0.81–0.86]

NaïveBayes

82.9 [80–86.2]

83.7 [75.7–92]

0.88 [0.83–0.92]

0.83 [0.8–0.86]

0.89 [0.86–0.91]

0.88 [0.83–0.92]

0.98 [0.97–0.99]

RandomForest

89.4 [87.4–91]

31.7 [23.3–40]

0.98 [0.96–1]

0.89 [0.87–0.91]

0.96 [0.94–0.98]

0.8 [0.74–0.86]

0.94 [0.92–0.96]

SMO

90 [88.4–91.6]

16.7 [16.7–16.7]

1 [1–1]

0.9 [0.88–0.92]

1 [1–1]

0.5 [0.5–0.5]

0.83 [0.81–0.86]

  1. ANN Artificial neural network (multilayer perceptron), AUC area-under-the-curve, PRC precision recall curve, ROC receiver operator characteristics, SMO sequential minimal optimization.
Back to article page