ML internals: Synthetic Minority Oversampling (SMOTE) Technique
Domino Data Lab
MAY 20, 2021
Machine Learning algorithms often need to handle highly-imbalanced datasets. Figure 3 shows visual explanation of how SMOTE generates synthetic observations in this case. A weighted nearest neighbor algorithm for learning with symbolic features. Machine Learning, 57–78. UCI machine learning repository.
136 
Let's personalize your content