Timeline for In elastic net regularisation, will dividing the OLS term the number of observations cause misleading results when cross-validating?
Current License: CC BY-SA 4.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 19, 2022 at 19:33 | history | edited | Richard Hardy | CC BY-SA 4.0 |
added 4 characters in body
|
| May 19, 2022 at 17:22 | comment | added | tobmo | @BenReiniger - agreed. My intuition is you'd much rather your regularisation strength diminishes as you add more data. I'm going to play around with some different implementations and see if I can prove this. | |
| May 19, 2022 at 17:21 | comment | added | tobmo | @Richard Hardy - no, sorry. I didn't even realise I'd done that! | |
| May 19, 2022 at 17:20 | history | edited | tobmo | CC BY-SA 4.0 |
edited body
|
| May 19, 2022 at 15:02 | history | edited | tobmo | CC BY-SA 4.0 |
deleted 2223 characters in body
|
| May 19, 2022 at 15:02 | comment | added | Ben Reiniger | Because the discrepancy boils down to "[the coefficients] from the sklearn model will be more strongly regularised", it seems worth a few experiments to see when/whether that's better. | |
| May 19, 2022 at 15:00 | history | edited | tobmo | CC BY-SA 4.0 |
deleted 2223 characters in body
|
| May 19, 2022 at 13:23 | history | edited | Sycorax♦ | CC BY-SA 4.0 |
deleted 4 characters in body
|
| May 19, 2022 at 13:10 | history | edited | Richard Hardy | CC BY-SA 4.0 |
added 1 character in body
|
| May 19, 2022 at 12:00 | history | tweeted | twitter.com/StackStats/status/1527257771464347649 | ||
| May 18, 2022 at 21:13 | history | edited | tobmo | CC BY-SA 4.0 |
added 1889 characters in body
|
| May 18, 2022 at 19:31 | answer | added | Sycorax♦ | timeline score: 3 | |
| S May 18, 2022 at 19:23 | review | First questions | |||
| May 18, 2022 at 20:15 | |||||
| S May 18, 2022 at 19:23 | history | asked | tobmo | CC BY-SA 4.0 |