Figure 2 | Scientific Reports

Figure 2

From: Imbalance-Aware Machine Learning for Predicting Rare and Common Disease-Associated Non-Coding Variants

Figure 2

Graphical representation of the SMOTE algorithm. (a) SMOTE starts from a set of positive (green points) and negative (blue points) examples; (b) It then selects a positive example (black) and its k nearest neighbors among the positives (yellow points, with k = 3), (c) Finally one of the k nearest neighbours is randomly selected (brown point) and a new synthetic positive example is added, by randomly generating an example (red point) along the straight line that connects the black and brown points. The procedure depicted in (b,c) is repeated for all the positives, by adding each time a new synthetic example similar (in an Euclidean sense) to the other positive examples.

Back to article page