Timeline for Random sampling without replacement [closed]
Current License: CC BY-SA 3.0
30 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 4, 2020 at 5:45 | history | closed |
The Fifth Marshal rydwolf Wheat Wizard♦ |
Needs details or clarity | |
| Dec 4, 2020 at 5:10 | review | Close votes | |||
| Dec 4, 2020 at 5:45 | |||||
| Nov 6, 2017 at 10:17 | answer | added | Arc676 | timeline score: 1 | |
| Nov 6, 2017 at 8:58 | answer | added | user58988 | timeline score: 0 | |
| Oct 18, 2017 at 15:37 | answer | added | mkst | timeline score: 0 | |
| Oct 18, 2017 at 13:10 | answer | added | Pavel Ruzankin | timeline score: 1 | |
| Jul 27, 2016 at 20:25 | answer | added | applejacks01 | timeline score: 0 | |
| Jul 27, 2016 at 13:30 | answer | added | Andreï V. Kostyrka | timeline score: 0 | |
| Jul 27, 2016 at 6:00 | history | edited | R. Kap | CC BY-SA 3.0 |
than Than THAN!
|
| Mar 7, 2012 at 11:04 | answer | added | Paolo | timeline score: 1 | |
| Mar 5, 2012 at 12:37 | answer | added | sinedcm | timeline score: 0 | |
| Feb 15, 2012 at 21:16 | answer | added | steenslag | timeline score: 1 | |
| Feb 4, 2012 at 10:13 | answer | added | celtschk | timeline score: 0 | |
| Feb 2, 2012 at 12:28 | answer | added | Hauleth | timeline score: 0 | |
| Feb 2, 2012 at 2:45 | answer | added | hammar | timeline score: 6 | |
| Feb 2, 2012 at 0:05 | answer | added | Samuel Duclos | timeline score: 0 | |
| Jan 31, 2012 at 3:33 | answer | added | user unknown | timeline score: 0 | |
| Jan 31, 2012 at 3:17 | comment | added | user unknown | And since it is a performance hack, the question is important, what type of test is performed. For example, if you have to pick 10,000 elements of 10,001, it would be fast to shuffle the 10,001 and pick the first 10,000. But if you just have to pick 1,000 of 100,000 you can put randomly chosen elements into a set, until it is filled with 1,000 elements (which is my solution so far), but this will be slow, if you have most of the elements of a very big collection. | |
| Jan 31, 2012 at 3:12 | comment | added | user unknown | If the request is to pick 5 distinct values in the range of (0..3) - shall we sanitize against such unfulfillable requests? | |
| Jan 30, 2012 at 19:16 | answer | added | AShelly | timeline score: 1 | |
| Jan 28, 2012 at 14:26 | history | tweeted | twitter.com/#!/StackCodeGolf/status/163266586805940224 | ||
| Jan 28, 2012 at 5:11 | history | edited | Jim McKeeth | CC BY-SA 3.0 |
added 269 characters in body
|
| Jan 28, 2012 at 5:08 | comment | added | Jim McKeeth | @IlmariKaronen: RE: Randomness: I've seen implementations before that were woefully unrandom. Either they had a heavy bias, or lacked the ability to produce different results on consecutive runs. So we are not talking cryptographic level randomness, but more random than the Accounting Department's random number generator in Dilbert. | |
| Jan 27, 2012 at 20:34 | answer | added | Blazer | timeline score: 1 | |
| Jan 27, 2012 at 11:12 | comment | added | Joey | Ah, indeed. But I'm not sure whether there are well-established libraries and tools for measuring randomness of sampling :-) | |
| Jan 27, 2012 at 11:11 | history | edited | Peter Taylor | CC BY-SA 3.0 |
Tighten spec
|
| Jan 27, 2012 at 11:07 | comment | added | Peter Taylor | @Joey, sure it is. It's random sampling without replacement. As long as no-one claims that the different positions in the list are independent random variables there's no problem. | |
| Jan 27, 2012 at 8:36 | comment | added | Joey | TestU01 is probably a bit harsh, I guess. Does criterion 3 imply a uniform distribution? Also, why the non-repeating requirement? That's not particularly random, then. | |
| Jan 27, 2012 at 6:35 | comment | added | Ilmari Karonen | Is the output supposed to pass something like the DIEHARD or TestU01 tests, or how will you judge its randomness? Oh, and should the code run in 32 or 64 bit mode? (That will make a big difference for optimization.) | |
| Jan 27, 2012 at 0:01 | history | asked | Jim McKeeth | CC BY-SA 3.0 |