Timeline for Is it possible to obtain orthogonal (but not normalized) vectors after QR factorization?
Current License: CC BY-SA 4.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 17, 2021 at 7:24 | comment | added | Federico Poloni | @M.Farooq Formulas (6.6) and (8.3). They're not applied to a vector, but (8.3) is basically the operator that I suggested. | |
| Dec 16, 2021 at 23:47 | comment | added | ACR | I got hold of Trefethen-Bau's book but I could not find properties of $QQ^T$ times a vector. The other $Q^TQ$ is discussed at length. | |
| Dec 16, 2021 at 14:04 | comment | added | ACR | @FedericoPoloni, Thanks. | |
| Dec 16, 2021 at 7:14 | comment | added | Federico Poloni | These are properties of orthogonal projections; they are treated for instance in the Trefethen-Bau book. | |
| Dec 15, 2021 at 18:56 | comment | added | ACR | Do you remember a reference which shows the above property m-QQ'm? | |
| Dec 15, 2021 at 18:55 | vote | accept | ACR | ||
| Dec 15, 2021 at 18:55 | comment | added | ACR | @FedericoPoloni, Thank you. | |
| Dec 15, 2021 at 18:23 | comment | added | Federico Poloni |
[email protected] Use the thin QR, [Q, R] = qr(M, 0), and associate from the right, Q*(Q'*m).
|
|
| Dec 15, 2021 at 17:22 | comment | added | ACR | I tried a very small matrix, and you are right it works, but with large data, say, 80000 points, MATLAB is giving a memory error of 47 GB requirement. | |
| Dec 15, 2021 at 16:59 | comment | added | ACR | @Federico, I am getting a memory error in MATLAB for Q*Q' where the prime is the transpose. My actual data is several thousand points and it says 47 GB memory required. Q is unfortunately 80000x5. | |
| Dec 15, 2021 at 16:18 | review | Low quality posts | |||
| Dec 15, 2021 at 17:27 | |||||
| Dec 15, 2021 at 15:58 | history | answered | Federico Poloni | CC BY-SA 4.0 |