grseb9s / matlab-smote

Matlab implementation of Synthetic Minority Oversampling Technique (SMOTE)

MATLAB-SMOTE

Matlab implementation of Synthetic Minority Oversampling Technique (SMOTE)

A technique to generate synthetic samples as given by Chawla et. al., 2002

Usage:

X_smote = mySMOTE(X, N, k)

Inputs:

• X: Original dataset (n x d) matrix
• N: Percentage of data-augmentation intended, Typically, N > 100, if N < 100, then N is set to 100. (scalar)
• k: number of nearest neighbors to consider while performing augmentation (scalar)

Outputs:

• X_smote: augmented dataset containing original data as well. (m x d) matrix, where m > n