Figure 1 : Comparison of nearest neighbor and label propagation strategies.

From: Label Propagation Prediction of Drug-Drug Interactions Based on Clinical Side Effects

Figure 1

(a) toy training data with two known drugs (in green) interacting with the input; (b) nearest neighbor method only searches drugs similar to the training drugs (in yellow); (c) label propagation method searches the whole drug similarity graph, i.e., search yellow drugs’ nearest neighbors as well (in blue).