Skip to main content

Table 2 The prediction performances of all machine learning models.

From: Predicting sulfotyrosine sites using the random forest algorithm with significantly improved prediction accuracy

  10 20 30
  Spe Sen Tot AUR Spe Sen Tot AUR Spe Sen Tot AUR
LDA 70 78 73 0.80 76 89 80 0.87 82 83 82 0.88
QDA 85 50 73 0.82 88 44 73 0.80 91 72 84 0.84
CART 91 72 84 n.a. 76 83 78 n.a. 88 83 86 n.a.
1NN 91 72 84 n.a. 88 72 82 n.a. 85 72 80 n.a.
3NN 85 78 82 n.a. 94 72 86 n.a. 94 72 86 n.a.
5NN 94 72 86 n.a. 97 72 88 n.a. 97 72 88 n.a.
7NN 88 67 78 n.a. 88 78 84 n.a. 94 78 88 n.a.
9NN 94 50 78 n.a. 88 78 86 n.a. 94 78 88 n.a.
RF 97 83 92 0.93 97 83 92 0.95 97 83 92 0.94
ANN5 94 28 71 0.81 88 67 80 0.86 88 78 84 0.92
ANN10 100 33 76 0.82 94 78 88 0.94 91 72 84 0.92
ANN15 91 56 78 0.86 97 67 86 0.89 94 78 88 0.93
ANN20 91 56 78 0.88 94 72 86 0.96 94 78 88 0.93
SVM 87 78 83 0.89 100 72 90 0.94 94 78 88 0.92