Timeline for Matching algorithm, match pairs: what would be the right way to do it?
Current License: CC BY-SA 3.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 20, 2017 at 8:51 | vote | accept | eyjin | ||
| Jan 18, 2017 at 15:51 | history | edited | eyjin | CC BY-SA 3.0 |
edited title, added information about the use
|
| Jan 17, 2017 at 17:52 | comment | added | David Richerby | @user3853544 It's not stable marriage; just maximum-weight matching (or maybe just matching). Stable marriage requires preference relations and requires the graph to be bipartite. | |
| Jan 17, 2017 at 17:49 | answer | added | David Richerby | timeline score: 2 | |
| Jan 17, 2017 at 16:45 | comment | added | greybeard |
Maybe you can see that it is just not clear for me If you (who else would you expect to?) can't pin down how to define a "distance" (sum off absolute differences/squares are somewhat common), try to describe its use: what is this grouping good for?
|
|
| Jan 17, 2017 at 14:59 | comment | added | user12859 | This seems to be maximum-weight matching. Is it the entries or the rows or the columns that might be outliers? (i.e., which of those are what might not be considered?) | |
| Jan 17, 2017 at 13:53 | comment | added | user3853544 | I don't think this is a sort. Since your just pairing I think the Stable Marriage Problem can be used to solve this. en.wikipedia.org/wiki/Stable_marriage_problem | |
| Jan 17, 2017 at 13:21 | history | edited | eyjin | CC BY-SA 3.0 |
tried to rearrange the question for better understandability
|
| Jan 17, 2017 at 12:33 | review | Close votes | |||
| Jan 17, 2017 at 17:45 | |||||
| Jan 17, 2017 at 12:16 | comment | added | paparazzo | "But not just based on the average of each row, but rather in dependency of the values to one another." Is not clear. | |
| Jan 17, 2017 at 12:04 | comment | added | greybeard | It is possible to create pairs and check validity. To sort by (decreasing) correlation, you need to define a measure. | |
| Jan 17, 2017 at 10:59 | comment | added | adrianN | If the efficiency of the algorithm is of no concern, just compare all possible pairs. | |
| Jan 17, 2017 at 10:52 | review | First posts | |||
| Jan 17, 2017 at 12:01 | |||||
| Jan 17, 2017 at 10:51 | history | asked | eyjin | CC BY-SA 3.0 |