Re: QRDecomposition

[Cleve Moler <moler@mathworks.com>]

On Thu, 2 Nov 2000, King, Jeffrey D wrote:

> 
> Perusing the archived e-mail I noticed the problem of the mxn vs mxm Q matrix from the decomposition.  I am created a Java application from a working matlab solution which calls lscov.  Inside lscov matlab uses the full decomposition.  Any way I can get that from your product?
> 
> Any help would be appreciated.
> 
> Thanx,
> 
> Jeff King
> The Boeing Company
> 
> 

Hi --

Yes, the algorithm used by MATLAB's "lscov" needs the full QR decomposition.
That's not available from the current version of JAMA.  You (or I) would
have to modify the Java code to give the full Q.

  -- Cleve Moler