| Re: [eigen] Error: trying to recover Q matrix in sparse QR factorization |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen <eigen@xxxxxxxxxxxxxxxxxxx>
- Subject: Re: [eigen] Error: trying to recover Q matrix in sparse QR factorization
- From: Gael Guennebaud <gael.guennebaud@xxxxxxxxx>
- Date: Tue, 7 May 2013 15:36:16 +0200
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=x-received:mime-version:in-reply-to:references:from:date:message-id :subject:to:content-type:content-transfer-encoding; bh=3IvO2L8oVDHkerP8FJEdSN/5zTaDPRDiGjhgWxi1QxQ=; b=rnLsvurot/ZFWMQpQZORyQN1gz++QETLuzX9FleVoZd5dNhfijlb96J5/MGd1jdrKo 9/scQR57bAUjEgsKoxiceN1TupE1gWiOyjSNHScvd8CaYAmj9F/jcfOytA9T1bXGJ4O9 OrroURFMZwKMw47+3CZLLVBvkoHqODijjghCItmdS98Rv3a1ruZvENgg2yUa5jeOUq5D MXr6Jld5jt6AWOnNdlmPEbdLspHp/WXXYX3PTu+HX3d8V+vs32H+xZToqW6Gv002hcms S7BXLHWd8js9NX4FA3rFgXNerk5PljuPx6BpGv9b7XU0RjdRZxBx3ITD1/GLB2OXfKam 4k9g==
Need to be fixed: http://eigen.tuxfamily.org/bz/show_bug.cgi?id=596
In the meantime, multiplying by the identity matrix is the way to go.
gael
On Tue, May 7, 2013 at 2:58 PM, Christoph Hertzberg
<chtz@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:
> On 07.05.2013 05:03, Avneesh Saluja wrote:
>>
>> Hello,
>>
>> I have a question regarding the sparse QR implementation in the latest
>> development build of Eigen.
>>
>> I would like to recover the Q matrix after the factorization, but can't
>> quite do so. I have followed the suggestion in the following link to get
>> the SparseMatrix representation of Q (where I assume A is the original
>> matrix that we are taking the factorization of):
>>
>> http://eigen.tuxfamily.org/dox-devel/classEigen_1_1SparseQR.html#ac27067ab301641fe6cc35c216a9b18a2
>
>
> That looks mostly like a documentation issue. In general, it does not make
> sense to store Q as a sparse matrix, since usually it is not anything near
> sparse. You can directly use qr_factorizer.matrixQ() inside expressions. If
> you really need the actual Q, I guess the simplest way would be to multiply
> by an identity matrix.
>
>
> Christoph
>
>
> --
> ----------------------------------------------
> Dipl.-Inf., Dipl.-Math. Christoph Hertzberg
> Cartesium 0.049
> Universität Bremen
> Enrique-Schmidt-Straße 5
> 28359 Bremen
>
> Tel: +49 (421) 218-64252
> ----------------------------------------------
>
>