help-gsl
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-gsl] Thin QR decomposition


From: Patrick Alken
Subject: Re: [Help-gsl] Thin QR decomposition
Date: Mon, 9 Oct 2017 09:06:22 -0600
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.3.0

On 10/09/2017 06:03 AM, Ines Arous wrote:
> Hi GSL team,
>
> I am trying to use the QR function that is provided by GSL (which is very 
> effective compared to other QR implementation I found online). My problem is 
> that it computes the full decomposition while I need the thin QR 
> decomposition.
> Is there a way to compute the thin QR decomposition using GSL? If so can you 
> please point me on how I can do that?
>
> Best regards,
> Thanks!
> Inès Arous

What exactly do you need to do? Are you trying to form the explicit Q1
and Q2 matrices? You almost never need to do this.

If you just want the product Q^T b, you can use gsl_linalg_QR_QTvec().
If you want to compute the residual norm || y - A*x ||, this can be done
without explicitly forming Q1 and Q2. 

If you really do need Q1, Q2 and R(NxN), you can use
gsl_linalg_QR_unpack() and just take submatrix views to get each factor.

Patrick




reply via email to

[Prev in Thread] Current Thread [Next in Thread]