Imbalanced data

Hello, I have more than 1500 X train data, with more than 500 labels (Y). There are imbalanced data, because It contain small talks intent which are just provide few examples but for commerce features we have more than 15 examples for each intent. What type of model can handle this type of data? Are use tf word embedding will get nice results?

Thank you, any comment appreciate :slight_smile:

Hi @khaerulumam42 what kind of intents are these? 500 intents with only 1500 examples seems like too many intents and not enough examples