Skip to main content

Table 2 AUC performance in the cross-validation experiment on the external validation (ChEMBL 19) dataset

From: Accurate and efficient target prediction using a potency-sensitive influence-relevance voter

Cutoff (\(\upmu \hbox {M}\)) PS-IRV SVM RF
All datasets
 \(1\) 0.70 0.69 0.68
 \(5\) 0.69 0.67 0.67
 \(10\) 0.69 0.66 0.67
Datasets with more than 100 molecules
 \(1\) 0.71 0.70 0.70
 \(5\) 0.70 0.68 0.69
 \(10\) 0.70 0.67 0.67
Datasets with more than 200 molecules
 \(1\) 0.72 0.72 0.71
 \(5\) 0.71 0.69 0.70
 \(10\) 0.70 0.68 0.68
  1. Models were trained on the ChEMBL 13 dataset
  2. Each section of the table shows the average performance for datasets of different sizes
  3. Best results within each group are in italics