Timeline for How to project a new vector onto the PC space using kernel PCA?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 1, 2016 at 23:22 | history | edited | amoeba | CC BY-SA 3.0 |
slightly shortened the title
|
| Dec 3, 2014 at 16:59 | vote | accept | giuseppe | ||
| Dec 2, 2014 at 23:49 | answer | added | amoeba | timeline score: 13 | |
| Dec 2, 2014 at 16:32 | history | edited | amoeba | CC BY-SA 3.0 |
light editing
|
| Dec 1, 2014 at 16:17 | comment | added | giuseppe | Sorry, you have reason, it was a typo error. N is the number of samples. Nevertheless I found a MATLAB function that performs what I need. From what I understood, it is possible to compute a new Kernel matrix that computes the kernel function between training data and new data. Successively it is possible to project this new kernel matrix in the same way obtaining the projection of the new data. | |
| Dec 1, 2014 at 16:13 | history | edited | giuseppe | CC BY-SA 3.0 |
deleted 1 character in body
|
| Nov 30, 2014 at 12:50 | history | edited | giuseppe | CC BY-SA 3.0 |
Improved formatting and question
|
| Nov 30, 2014 at 12:16 | comment | added | ttnphns | Exactly as you projected the existent data. Assuming you won't undertake re-doing the dimensionality reduction (estimation) part of the analysis with every new point coming. | |
| Nov 30, 2014 at 11:47 | history | asked | giuseppe | CC BY-SA 3.0 |