# Ambiguous [combination] tag

See (113 threads). Some of these are combinatorics questions about counting combinations, some are about linear combinations, combining distributions, combining this, combining that... The tag has no wiki. I don't think this tag is very useful in any of its meanings.

### Proposed way forward

• Those Qs that are about combinatorics should be retagged [combinatorics] or if both tags are present, the [combination] tag simply deleted. I don't think it is relevant to separate combinatorics questions that are about "combinations" from other combinatorics questions by tags.
• In other cases, the tag [combination] should on a case-by-case basis be either just deleted or retagged to something more descriptive if there is something tag-worthy, say, [combining-forecasts].
• +1. Would you be willing to do the job? – amoeba Aug 16 '16 at 12:23
• Not willing to commit any significant time but if this is agreed upon, I may process some questions every now and then – Juho Kokkala Aug 16 '16 at 12:32
• While going through the meta-analysis thread removing tags to do with weighting I noticed some questions tagged [combination]. This is tautologous, what is m-a after all except combination? When I get time I will delete the tags unless you get there first. – mdewey Aug 16 '16 at 15:05
• There is also a stray tag [combinatorial] with three threads. – amoeba Aug 16 '16 at 19:14
• All 3 [combinatorial]-questions seem to be about combinatorial optimization, perhaps that tag should be renamed combinatorial-optimization – Juho Kokkala Aug 16 '16 at 19:24
• I supplied a tag wiki excerpt saying DO NOT USE THIS TAG: it is ambiguous and has no meaning. Use more specific tags instead (usually [combinatorics]). – amoeba Aug 18 '16 at 23:08
• I have removed all (the handful) of [combination] tags from meta-analysis. Note that there is a tag [combining-p-values] which is distinct and valuable. – mdewey Aug 19 '16 at 14:47
• Somewhat related issue: There is a permutation tag (261 threads) which is mostly used to refer to permutation-based hypothesis tests, but sometimes refers to combinatoric problems related to permutations. Perhaps it would make sense to remove it from the latter ones (replacing with [combinatorics]) and then rename the tag to [permutation-test]. I realize it is a separate issue, we can raise it on Meta once this one is resolved. – amoeba Aug 20 '16 at 12:58
• @amoeba There would be some synergy benefits from deciding the fate of [permutation] now, however. While going through [combination] one could either i) sometimes kill two tags with one stone or ii) sometimes add [permutation], depending on what was decided. – Juho Kokkala Aug 20 '16 at 17:12
• My own opinion is that [permutation] is as superfluous as [combination]; questions about permutations and combinations can be tagged with [combinatorics]. Questions about permutation tests should be tagged with [permutation-test]. I might raise it on Meta at some point, but for now I would simply remove [permutation] from combinatorics threads that you are editing. – amoeba Aug 21 '16 at 22:07
• Question posted: meta.stats.stackexchange.com/questions/4280. – amoeba Aug 22 '16 at 13:47
• @mdewey Thanks for your continuing effort in eliminating the [combination] tag. Don't forget to post an answer here announcing that this task is over once you are done. – amoeba Sep 23 '16 at 12:35
• @amoeba I am perhaps over-reluctant just to delete [combination] without a replacement. Perhaps I should be more ruthless. – mdewey Sep 23 '16 at 13:12
• Juho, please see the updates to @mdewey's answer. [Combination] tag does not exist anymore, and the small sister tags have been cleaned up as well. Consider accepting his answer to mark this issue as resolved. – amoeba Nov 8 '16 at 12:00
• @amoeba Answer accepted, thanks for notifying me – Juho Kokkala Nov 8 '16 at 18:06

This tag is now empty, and if I understand correctly will auto-delete. there remain a few with the tag linear-combination which as @amoeba pointed out could apply to most posts here so unless anyone objects in comments here I will deal with them in a few days unless anyone else like to do that.

Update: I have removed the [linear-combination] tag from all 8 which had it and in one case added [contrasts]. I think [combining-models] could be replaced with [model-averaging] without loss of information and perhaps some of [combining-estimates] and [combining-predictions] could be similarly re-tagged. The one which stumps me is [combinatorial]. I think it means something different from [combinatorics] but I do not really understand the area well. If nobody objects here I will do the re-tagging I suggest and hope someone else will pick up the [combinatorial] issue, perhaps by writing a wiki excerpt.

Final, I hope, update: I have now removed all of [combining-predictions] and [combining-estimates] so when the tag fairy sprinkles her magic dust on them the last one will disappear.

• Great job. There is a couple of other minor tags starting with "combin", many of them without excerpts and of unclear usefulness. [combinatorial] with 7 Qs and no excerpt; [combining-predictions] with 5 Qs and no excerpt; [combining-estimates] with 4 Qs; [combining-models] with 1 Q and no excerpt. I did not look at them, but I have a feeling most of them are not useful. What do you think? – amoeba Sep 25 '16 at 21:38
• @amoeba I will have a look and perhaps report back. – mdewey Sep 26 '16 at 15:52
• Have you had a chance to take a look? Somebody just posted a suggestion on Meta to delete [combining-models], so perhaps we should finish up this tidying... – amoeba Oct 4 '16 at 17:08
• @amoeba yes, I have been having a siesta. I will have a look tomorrow. – mdewey Oct 4 '16 at 20:00
• @amoeba what is the protocol for this? Do I need to post another question or, for instance, is it OK just to remove the tag from [linear-combination] which would not really affect any of them? – mdewey Oct 5 '16 at 17:09
• With [linear-combination] you can definitely go ahead already now (nobody objected since Sep 25, see your answer). With other mentioned tags, I'd say you can also go ahead; or perhaps edit your answer to announce what you are going to do unless nobody objects. This is not worth a separate thread. – amoeba Oct 5 '16 at 19:10
• @amoeba I have now removed [linear-combination] from all 8 and in one case replaced it with [contrasts] – mdewey Oct 6 '16 at 15:33
• (1) Consider updating your answer accordingly. (2) Do you have any opinion on the other combining- tags that I listed in the first comment above? – amoeba Oct 27 '16 at 20:57
• [combinatorial] is no more. – amoeba Oct 29 '16 at 20:22
• Now removed [combining-models] in one case replacing with [model-averaging] – mdewey Oct 30 '16 at 15:11
• Thanks for taking care of it, @mdewey. – amoeba Nov 9 '16 at 14:16