Abstract: The long-tailed distribution of data poses a significant challenge in text classification tasks. The imbalanced distribution of samples among categories often hinders effective ...