Table 6 Performance of machine learning–based classifications of anti-ARS-antibodies.

From: Texture analysis of muscle MRI: machine learning-based classifications in idiopathic inflammatory myopathies

 

Accuracy

Sensitivity

Specificity

F-measure

AUC

(%)

(%)

(%)

LDA (TexF3 + TexF6 + TexF10)

78.3 ± 2.2

75.5 ± 3.3

81.5 ± 2.5

0.815 ± 0.025

0.792 ± 0.018

QDA (TexF3 + TexF5 + TexF10)

71.1 ± 2.9

68.3 ± 4.7

74.3 ± 3.2

0.715 ± 0.033

0.731 ± 0.030

SVM (TexF4 + TexF7 + TexF10)

76.4 ± 2.2

74.5 ± 4.0

78.7 ± 2.6

0.771 ± 0.025

0.752 ± 0.015

k-NN (TexF4 + TexF8 + TexF10)

73.2 ± 2.6

75.6 ± 2.6

70.5 ± 5.1

0.751 ± 0.022

0.742 ± 0.019

RF (TexF3 + TexF4 + TexF10)

65.8 ± 3.5

69.1 ± 5.2

62.0 ± 4.2

0.682 ± 0.037

0.692 ± 0.022

MLP (TexF4 + TexF9 + TexF10)

68.9 ± 3.9

72.6 ± 5.1

64.7 ± 6.4

0.713 ± 0.037

0.693 ± 0.029

  1. Note Data are means ± standard deviations. Feature name codes are as follows: TexF1 = kurtosis, TexF2 = interquartile range, TexF3 = total energy, TexF4 = cluster prominence, TexF5 = correlation, TexF6 = difference average, TexF7 = imc2, TexF8 = maximum probability, TexF9 = large dependence high gray-level emphasis, TexF10 = dependence non-uniformity, TexF11 = coarseness.
  2. LDA linear discriminant analysis, QDA quadratic discriminant analysis, SVM support vector machine, k-NN k-nearest neighbors classifier, RF random forest classifier, MLP multi-layer perceptron.
Back to article page