Timeline for SMOTE sampling in caret package in R
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 23 at 15:03 | comment | added | Meo | Thank you! I read your post regarding Classification vs Prediction. fharrell.com/post/classification/…. | |
| Oct 20 at 12:02 | comment | added | Frank Harrell | The short answer is don’t use any forced-choice classification on categorical Y that doesn’t naturally have balance. Instead model tendencies, i.e. probabilities of Y. | |
| Oct 19 at 14:23 | comment | added | Meo | I did. But I couldn’t find a discussion for using decision tree on censored outcome with imbalanced data. Can you please help me or guide me to the discussion? Thank you! | |
| Oct 19 at 12:31 | comment | added | Frank Harrell | Do you not search this site first when posting questions? This has been covered numerous times | |
| Oct 19 at 6:14 | comment | added | Meo | Can you please elaborate on your response? I am trying to build a classification tree to predict mortality. I used SMOTE to balance my dataset, but don’t know how to take care of the censoring problem since each individual in the data has different follow up time. That’s why I thought of using weights. I would really appreciate any advice. Thank you! | |
| Oct 17 at 14:01 | comment | added | Frank Harrell | This represents really problematic statistical practice. | |
| Oct 17 at 1:51 | history | answered | Meo | CC BY-SA 4.0 |