R devel - Jul 2023 - Concerns with SVD

Respected Development Team,

This is Durga Prasad reaching out to you regarding an issue/concern related
to Singular Value Decomposition SVD in R software package. I am attaching a
detailed attachment with this letter which depicts real issues with SVD in
R.

To reach the concern the expressions for the exponential of a matrix using
SVD and
projection tensors are obtained from series expansion. However, numerical
inconsistency is observed between the exponential of matrix obtained using
the function(svd()) used in R software.

However, it is observed that most of the researchers fraternity is engaged
in utilising R software for their research purposes and to the extent of my
understanding such an error in SVD in R software might raise the concern
about authenticity of the simulation results produced and published by
researchers across the globe.

Further, I am very sure that the R software development team is well versed
with the happening and they have any specific and resilient reasons for
doing so. I would request you kindly, to guide me through the concern.

Thank you very much.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: Concerns with SVD.pdf
Type: application/pdf
Size: 128838 bytes
Desc: not available
URL:
<https://stat.ethz.ch/pipermail/r-devel/attachments/20230716/3fad35ca/attachment.pdf>

Better check your definitions of SVD -- there are several forms, but all I
am aware of (and I wrote a couple of the codes in the early 1970s for the SVD)
have positive singular values.

JN


On 2023-07-16 02:01, Durga Prasad G me14d059 wrote:
> Respected Development Team,
> 
> This is Durga Prasad reaching out to you regarding an issue/concern related
> to Singular Value Decomposition SVD in R software package. I am attaching a
> detailed attachment with this letter which depicts real issues with SVD in
> R.
> 
> To reach the concern the expressions for the exponential of a matrix using
> SVD and
> projection tensors are obtained from series expansion. However, numerical
> inconsistency is observed between the exponential of matrix obtained using
> the function(svd()) used in R software.
> 
> However, it is observed that most of the researchers fraternity is engaged
> in utilising R software for their research purposes and to the extent of my
> understanding such an error in SVD in R software might raise the concern
> about authenticity of the simulation results produced and published by
> researchers across the globe.
> 
> Further, I am very sure that the R software development team is well versed
> with the happening and they have any specific and resilient reasons for
> doing so. I would request you kindly, to guide me through the concern.
> 
> Thank you very much.
> 
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel