Remove 2002 Remove Knowledge Discovery Remove Strategy Remove Testing
article thumbnail

ML internals: Synthetic Minority Oversampling (SMOTE) Technique

Domino Data Lab

In their 2002 paper Chawla et al. propose a different strategy where the minority class is over-sampled by generating synthetic examples. 2002) have performed a comprehensive evaluation of the impact of SMOTE- based up-sampling. Their tests are performed using C4.5-generated Generation of artificial examples.