I'm trying to apply xgboost and random forest for over and under sampling <br/>
for imbalance data<br/>

train shape -> (199991, 23)

[![enter image description here][1]][1]


  [1]: https://i.sstatic.net/efq46.png

but reverse my expectation. <br/>
the accuracy is went down for both case<br/>

1. is under and over sampling always not good? <br/>
2. what else do I have to consider when I apply over and under sampling?<br/>



1. xgboost randomforest
2. under over sampling
3. multi calssification 
4. imbalanced dataset