Several methodologies exist for oversampling, each with its own advantages and disadvantages. Some common techniques include:
Random Oversampling: This involves duplicating instances of the minority class until the dataset is balanced. While simple, this method can lead to overfitting. SMOTE (Synthetic Minority Over-sampling Technique): This technique generates synthetic samples by interpolating between existing minority samples. It helps in reducing overfitting compared to random oversampling. ADASYN (Adaptive Synthetic Sampling): An extension of SMOTE, ADASYN focuses on generating more synthetic samples for minority class instances that are harder to classify.
