[
https://issues.apache.org/jira/browse/MATH-665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13100350#comment-13100350
]
Sébastien Brisard edited comment on MATH-665 at 9/8/11 2:24 PM:
----------------------------------------------------------------
I've looked into NR and [SLATEC|http://www.netlib.org/slatec/lin/dqrsl.f] (see
above line 150). In both cases a similar test for strict equality is carried
out. This is strange indeed; what would be a reasonable threshold?
was (Author: celestin):
I've looked into NR and [SLATEC|http://www.netlib.org/slatec/lin/dqrsl.f].
In both cases a similar test for strict equality is carried out. This is
strange indeed; what would be a reasonable threshold?
> QRDecomposition does not have a singularity threshold
> -----------------------------------------------------
>
> Key: MATH-665
> URL: https://issues.apache.org/jira/browse/MATH-665
> Project: Commons Math
> Issue Type: Bug
> Reporter: Phil Steitz
>
> QRDecompositionImpl tests elements of rDiag for exact equality to 0 in
> checking singularity. A singularity threshold should be defined for this
> class and used in the singularity test.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
