Timeline for How to choose a predictive model after k-fold cross-validation?
Current License: CC BY-SA 3.0
23 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 23, 2022 at 15:40 | answer | added | Fred | timeline score: 0 | |
| Oct 11, 2021 at 10:10 | answer | added | TheodorosNikoleris | timeline score: 0 | |
| Oct 21, 2018 at 13:39 | answer | added | Saeed | timeline score: 4 | |
| Jan 13, 2018 at 6:56 | answer | added | Patrick Ng | timeline score: 55 | |
| Apr 4, 2017 at 14:22 | history | edited | Berk U. | CC BY-SA 3.0 |
improved working
|
| Jan 4, 2017 at 7:40 | answer | added | Canhua Li | timeline score: 3 | |
| Sep 19, 2016 at 2:15 | history | edited | Andre Silva | CC BY-SA 3.0 |
removed tag as per meta post: http://meta.stats.stackexchange.com/questions/2804/what-should-the-best-practices-tag-be-for
|
| Mar 13, 2016 at 21:23 | history | edited | Berk U. | CC BY-SA 3.0 |
deleted 48 characters in body; edited title
|
| Feb 11, 2016 at 17:18 | history | edited | Berk U. | CC BY-SA 3.0 |
deleted 4 characters in body
|
| S Jun 17, 2015 at 20:12 | history | edited | AdamO |
Adding "model selection" tag to the question.
|
|
| S Jun 17, 2015 at 20:12 | history | suggested | Kingz |
Adding "model selection" tag to the question.
|
|
| Jun 17, 2015 at 20:02 | review | Suggested edits | |||
| S Jun 17, 2015 at 20:12 | |||||
| Nov 5, 2014 at 16:59 | answer | added | DataExplorer | timeline score: 1 | |
| Mar 15, 2013 at 20:36 | vote | accept | Berk U. | ||
| Mar 15, 2013 at 16:58 | comment | added | Frank Harrell | For 10-fold cv it is best to do $\geq 50$ repeats. More repeats will be needed with 5-fold. These are rules of thumb. A single 10-fold cv will given an unstable answer, i.e., repeat the 10 splits and you get enough of a different answer to worry. | |
| Mar 15, 2013 at 14:47 | comment | added | BGreene | @Frank Harrell, why do you say 100 repetitions is necessary (I usually use 10 reps on 10 fold), is this a rule of thumb as the OP didn't give any specifics? | |
| Mar 15, 2013 at 14:06 | history | edited | user88 | CC BY-SA 3.0 |
edited title
|
| Mar 15, 2013 at 13:40 | history | tweeted | twitter.com/#!/StackStats/status/312558816346648576 | ||
| Mar 15, 2013 at 12:33 | comment | added | Frank Harrell | You will need to repeat 5-fold CV 100 times and average the results to get sufficient precision. And the answer from @bogdanovist is spot on. You can get the same precision of accuracy estimate from the bootstrap with fewer model fits. | |
| Mar 15, 2013 at 12:22 | answer | added | cbeleites | timeline score: 46 | |
| Mar 15, 2013 at 11:13 | comment | added | BGreene | You might find the answers in the following question helpful: stats.stackexchange.com/questions/2306/… | |
| Mar 15, 2013 at 2:53 | answer | added | Bogdanovist | timeline score: 356 | |
| Mar 15, 2013 at 2:21 | history | asked | Berk U. | CC BY-SA 3.0 |