Fig. 2
From: Deciding when to stop: efficient experimentation to learn to predict drug-target interactions
Comparison of random sampling (green) to uncertainty sampling (black: random column initialization, red: diversity based initialization) on the four data sets: (a) Nuclear Receptor, (b) GPCR, (c) Ion Channel, (d) Enzyme. The solid and the dotted line represent respectively the mean and the standard deviation of 5 random initializations. For random sampling five random runs were performed for each initialization