Skip to main content

Table 3 Prediction accuracies obtained by increasing fractions of the 441 elements of a window of size 21

From: Prediction of the burial status of transmembrane residues of helical membrane proteins

Fraction used in the prediction Linear regression SVR C – 101 SVR C – 1 SVR C – 0.1 SVR C – 0.01
0.052 75.65 70.36 73.45 75.21 74.89
0.1 76.61 71.06 75.88 75.97 74.57
0.15 76.90 70.78 75.11 76.20 74.09
0.2 77.21 70.65 75.14 75.78 73.58
0.25 76.45 70.01 75.59 75.78 73.04
0.3 76.04 71.13 75.11 75.24 72.56
0.35 75.72 71.54 74.79 75.27 71.54
0.4 75.91 72.69 74.76 75.11 71.80
0.45 75.91 72.72 75.11 74.95 71.86
0.5 76.13 72.82 75.33 75.11 71.67
0.55 76.39 72.63 75.43 75.43 72.08
0.6 76.04 72.69 75.43 75.14 71.61
0.65 75.33 72.94 75.24 75.30 70.40
0.7 74.86 73.01 74.98 74.92 70.24
0.75 75.33 73.20 75.43 74.86 69.31
0.8 75.75 72.75 75.75 74.44 68.48
0.85 75.62 72.59 75.75 74.22 67.97
0.9 75.24 72.34 75.14 74.00 67.65
0.95 75.21 72.79 75.46 74.22 67.53
1.0 75.59 73.26 75.97 74.16 67.85
  1. 1Regularization constant C was set to 10.
  2. 2Meaning that the top 5% of the 441 elements when ranked by the Fisher's index were input for the prediction.