I have a highly unbalanced text classification data. I am trying to over-sample through SMOTE.
I have a doubt that applying SMOTE over sequence of word indices will give me valid data points or not (since SMOTE used nearest neighbor like method)?
Asked
Active
Viewed 1,284 times
1
-
https://datascience.stackexchange.com/a/27758/8878 – Kasra Manshaei Feb 13 '18 at 13:12