Is augmentation effective to improve prediction in imbalanced text datasets?

Re-assign community