Skip to main content

Table 1 The performance of SVM model (learning parameter: g: 0.1 c: 2 j: 3) using amino acid sequence (The SVM parameter g (in RBF kernel), c: parameter for trade-off between training error & margin, j: cost-factor)

From: Identification of ATP binding residues of a protein from its primary sequence

Thres Sen Spec Accuracy MCC
-1 100 1.73 50.87 0.09
-0.9 99.87 2.88 51.37 0.11
-0.8 99.67 4.39 52.03 0.13
-0.7 99.25 6.51 52.88 0.15
-0.6 98.36 10.31 54.34 0.18
-0.5 96.89 15.78 56.33 0.22
-0.4 93.75 23.54 58.64 0.24
-0.3 88.94 32.9 60.92 0.26
-0.2 83.4 43.31 63.36 0.29
0 65.53 66.97 66.25 0.33
0.1 54.6 76.99 65.79 0.32
0.2 43.11 84.98 64.04 0.31
0.3 33.94 91.16 62.55 0.31
0.4 25.7 94.7 60.2 0.28
0.5 18.07 97.15 57.61 0.25
0.6 12.64 98.53 55.58 0.22
0.7 8.71 99.18 53.94 0.19
0.8 6.22 99.54 52.88 0.16
0.9 3.93 99.77 51.85 0.13
1 1.87 99.84 50.85 0.08
  1. (Bold values indicate the point where sensitivity and specificity is roughly equal with maximum MCC.)