Timeline for What are good metrics to assess the quality of a PCA fit, in order to select the number of components?
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 22, 2014 at 15:25 | history | edited | amoeba | CC BY-SA 3.0 |
made the title more specific
|
| May 28, 2014 at 19:40 | history | tweeted | twitter.com/#!/StackStats/status/471737908672880640 | ||
| May 27, 2014 at 15:56 | comment | added | CloseToC | Since one of your tags is info theory: An indirect way of assessing whether PCA works is to check the assumptions under which information theory tells us it has low info loss for a given dimension reduction. Wiki says this is so when your data is a sum of gaussian signal plus gaussian noise. en.wikipedia.org/wiki/… | |
| May 27, 2014 at 15:51 | answer | added | Nikos M. | timeline score: 3 | |
| May 27, 2014 at 10:51 | vote | accept | bigTree | ||
| May 27, 2014 at 8:18 | answer | added | Deathkill14 | timeline score: 18 | |
| May 27, 2014 at 7:47 | comment | added | Stephan Kolassa | Strictly speaking, there is no "redundant" information, unless your initial data were perfectly collinear. One usually sees percentage of variance retained ("we used the first five principal components, which accounted for 90% of the variance"). I'm interested in seeing alternatives. | |
| May 27, 2014 at 7:40 | history | asked | bigTree | CC BY-SA 3.0 |