'would it be possible to combine Balancing techniques with Doc2vec

I have a dataset that is imbalanced and wanted to use techniques such as SMOTE, ADASYN etc, to balance it out.

Would it be acceptable to use Doc2vec and then incorporate SMOTE to the training sample?



Solution 1:[1]

The best way to know if SMOTE (or some other augmentation technique) might help with your particular data, goals, & classification-algorithms is to try it, and see if it improves results on your relevant evaluations, compared to not using it.

It's "acceptable" if it works; there's no other outside/1st-principles to judge its potential applicability, without trying it.

Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source
Solution 1 gojomo