What is the best way to use SMOTE Node
Before partitioning or after partitioning ?
In my opinion you should use it after partitioning and only for your training data. SMOTE is used to create more data points, but when evaluating your model quality you want that quality to be measured against the real data points, not made-up ones. You can also split off another validation set and use it to see if SMOTEing brings you any gains at all.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.