Skip to main content

Table 3 Predictive performance of structural features for the four benchmark datasets

From: Prediction of HIV-1 protease cleavage site using a combination of sequence, structural, and physicochemical features

Features DT LR ANN
Acc.(%) AUC Acc.(%) AUC Acc.(%) AUC
746 Dataset
 SSE 62.1 0.626 59.4 0.715 78.3 0.838
 SA 83.7 0.791 78.4 0.771 81.0 0.771
 Str_All 83.7 0.791 70.2 0.806 78.4 0.897 *
1625 Dataset
 SSE 81.7 0.756 76.8 0.673 85.3 0.742
 SA 91.4 0.920 89.0 0.961 96.3 0.977
 Str_All 91.5 0.920 85.4 0.936 89.0 0.935
Schilling Dataset
 SSE 87.1 0.500 88.3 0.775 88.3 0.800
 SA 89.5 0.788 84.0 0.828 87.1 0.840
 Str_All 89.5 0.788 83.4 0.824 85.8 0.843
Impens Dataset
 SSE 85.1 0.500 85.1 0.729 87.2 0.761
 SA 89.3 0.736 89.3 0.918 95.7 0.950
 Str_All 87.2 0.571 89.3 0.857 89.3 0.914
  1. *The best accuracy and AUC in each dataset are underlined