[
https://issues.apache.org/jira/browse/MATH-1100?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13907188#comment-13907188
]
alberto trivellato commented on MATH-1100:
------------------------------------------
Thank Luc for your fast reply.
Given the dimension of the matrix, there is a high probability of a loss
of precision in java double calculations, and so that small values
appear in the diagonal of the triangular factor.
I agree with your analysis and I will try to use the threshold.
Thank you again,
best regards
--
Alberto Trivellato
> QR factorization fails in revealing rank-deficient matrix
> ---------------------------------------------------------
>
> Key: MATH-1100
> URL: https://issues.apache.org/jira/browse/MATH-1100
> Project: Commons Math
> Issue Type: Bug
> Affects Versions: 3.2
> Environment: Java HotSpot(TM) Client VM (build 19.1-b02, mixed mode,
> sharing) an windows 7 professional
> Reporter: alberto trivellato
> Attachments: QRtest.zip
>
>
> given a matrix that has not full rank, the method getSolver().isNonSingular()
> of the class QRDecomposition returns true.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
