When SMOTE or ADASYN create synthetic sample and the corresponding label, is the quality of the label trust worthy OR should I should double check the label by myself ?

Particularly in a hypothetical situation of a heavy imbalance digit recognition datasets (10 000 image of digit 9 and 1000 image of digit 10) can it actually create 9000 synthetic sample of the minority digit 10 and correctly label it owns 9000 synthetic sample ?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.