Synthetic Oversampling: Theory and A Practical Approach Using LLMs to Address Data Imbalance

Ryumei Nakada, Yichen Xu, Lexin Li +1

Imbalanced classification and spurious correlation are common challenges in data science and machine learning. Both issues are linked to data imbalance, with certain groups of data…