Timeline for XGBoost Classifier not capturing extreme probabilities
Current License: CC BY-SA 4.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 8, 2022 at 22:46 | comment | added | usεr11852 | Please someone mention isotonic regression, Platt scaling and beta calibration. | |
| Mar 8, 2022 at 21:26 | answer | added | ljubomir | timeline score: 2 | |
| Mar 3, 2022 at 10:06 | comment | added | Lodinn | If you want to add some special consideration to these extreme outliers, consider transforming your loss function: as of now, if going from 0.97 to 0.99 means that the predictions between say 0.2 and 0.8 would shift ever so slightly off the target, the model might "think" of it as an overall deterioration - you might have a different opinion. Make your target function reflect what you actually want out of the model! | |
| Mar 3, 2022 at 3:03 | history | became hot network question | |||
| Mar 3, 2022 at 0:00 | history | tweeted | twitter.com/StackStats/status/1499172717077147648 | ||
| Mar 2, 2022 at 20:02 | answer | added | Tim | timeline score: 5 | |
| Mar 2, 2022 at 19:52 | answer | added | Sycorax♦ | timeline score: 7 | |
| Mar 2, 2022 at 19:20 | comment | added | dfried | @Sycorax edited. Thanks! One comment on features: Right now it's just time, score, and the teams' qualities. I'll add more eventually, but given these features alone, I would expect the model to pick up on the fact that given certain time-score combinations, there aren't any features that are going to drag the probability away from 0 or 1 | |
| Mar 2, 2022 at 19:19 | history | edited | dfried | CC BY-SA 4.0 |
added 936 characters in body
|
| Mar 2, 2022 at 19:13 | comment | added | Eli | Are 0.03 and 0.97 meaningly different from .0001 or .9999? Are you still classifying 0.03 as 0 and 0.97 as 1? If so, it's probably not worth examining further. | |
| Mar 2, 2022 at 19:12 | comment | added | Sycorax♦ | How do you know that the probability should be $10^{-4}$ instead of 0.03? What hyperparmeters are you using? How many observations, features, and members of each class do you have? What is your loss function? Please edit to clarify. | |
| S Mar 2, 2022 at 18:58 | review | First questions | |||
| Mar 2, 2022 at 19:13 | |||||
| S Mar 2, 2022 at 18:58 | history | asked | dfried | CC BY-SA 4.0 |