Timeline for Regression: How to deal with positive skewness in continuous target variable
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 26, 2022 at 17:04 | comment | added | datainsights | Since the target is the one that is skewed you could try under or oversampling. Another thing is that tree based models are not affected by skewness in data. You could also bin the target variable | |
| Jan 6, 2020 at 14:55 | comment | added | Peter | When I apply a square transformation, I have the problem, that negative predicted values cannot be re-transformed, since there is no real number solution for that. square root would work, but it delivers no real improvement. The predicted distribution look very much like the one posted in my question (bottom). | |
| Jan 6, 2020 at 3:49 | history | edited | Stephen Rauch♦ | CC BY-SA 4.0 |
Minor grammar / formatting
|
| Jan 6, 2020 at 3:40 | review | Low quality posts | |||
| Jan 6, 2020 at 3:49 | |||||
| Jan 6, 2020 at 3:25 | review | First posts | |||
| Jan 6, 2020 at 3:49 | |||||
| Jan 6, 2020 at 3:21 | history | answered | datainsights | CC BY-SA 4.0 |