Timeline for Interpreting plot of PCA results (from 3 to 2 dimensions)
Current License: CC BY-SA 4.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 3, 2023 at 7:15 | history | edited | David Marx | CC BY-SA 4.0 |
their -> they're
|
| Nov 21, 2013 at 23:26 | history | edited | Scortchi♦ | CC BY-SA 3.0 |
fixed typos
|
| Aug 5, 2013 at 4:40 | history | edited | David Marx | CC BY-SA 3.0 |
deleted 3 characters in body
|
| Aug 5, 2013 at 4:38 | comment | added | David Marx | Updated my answer to demonstrate why your plot looks the way it does. Short version: because X and Y are tightly correlated, PCA combines them into a single dimension to allow you to explore the relationship between those variables and Z. The X-Z and Y-Z relationships are effectively the same thing because X and Y are so tightly correlated. Your PCA transformation is showing you the much less structured (XY)-Z relationship. | |
| Aug 5, 2013 at 4:33 | history | edited | David Marx | CC BY-SA 3.0 |
added 645 characters in body
|
| Aug 5, 2013 at 4:27 | history | edited | David Marx | CC BY-SA 3.0 |
added 645 characters in body
|
| Aug 5, 2013 at 3:55 | history | edited | David Marx | CC BY-SA 3.0 |
added 1369 characters in body
|
| Aug 4, 2013 at 19:44 | comment | added | Nancy | Thanks for the quick response. I've read though the link you suggested several times-- agreed, it's a fantastic resource. I suspect that I'm over thinking my question. My concern is that x and y are very tightly correlated with plotted against each other, but the "flattened" plot looks like a bunch of random points. Is this just a result of having the z data being so uncorrelated? | |
| Aug 4, 2013 at 19:22 | history | answered | David Marx | CC BY-SA 3.0 |