What are the differences between oversampling and undersampling in data processing?

What are the differences between oversampling and undersampling in data processing?

Oversampling involves increasing the number of instances in the minority class, while undersampling reduces instances in the majority class.

Can you give me examples of when to use each technique?

Use oversampling when you have a small minority class and want to improve model performance. Use undersampling when the majority class is too large and may lead to overfitting.

What are the potential drawbacks of each method?

Oversampling can lead to overfitting, while undersampling may result in loss of important data.

Are there any hybrid methods that combine both techniques?

Yes, hybrid methods like SMOTE (Synthetic Minority Over-sampling Technique) create synthetic examples while also considering the majority class.