Skip to main content

Table 8 10-folds cross-validation predictive accuracies for Aleph, ProGolem and SVM

From: Automated identification of protein-ligand interaction features using Inductive Logic Programming: a hexose binding case study

  Learning algorithm
Fold Aleph 1 ProG. 1 Aleph 2 ProG. 2 SVM
1 50.0% 75.0% 56.3% 75.0% 81.3%
2 68.8% 81.3% 68.8% 81.3% 87.5%
3 62.5% 68.8% 68.8% 93.8% 87.5%
4 50.0% 56.3% 68.8% 75.0% 75.0%
5 75.0% 81.3% 56.3% 81.3% 75.0%
6 68.8% 87.5% 81.3% 87.5% 87.5%
7 75.0% 81.3% 75.0% 81.3% 93.8%
8 93.8% 81.3% 75.0% 93.8% 87.5%
9 68.8% 75.0% 75.0% 81.3% 75.0%
10 56.3% 56.3% 87.5% 81.3% 62.5%
Mean 66.9% 74.4% 71.3% 83.2% 81.3%
Std Dev 13.2% 10.8% 9.8% 6.6% 9.3%
  1. The 1 besides Aleph and ProGolem stands for the atom-only representation and the 2 for the amino acid representation. SVM uses a different representation (see text).