Re: pseudoinverse matrix

From: Brian Borchers <borchers.brian@xxxxxxxxx>
Date: Tue, 18 Nov 2008 18:32:17 -0800 (PST)

On Nov 18, 10:25 am, Tim De Meyer

The text sais:
"Calculate a weight vector, w, that makes X * w = tnew. This is done by a PLS estimation giving a generalised inverse = Xinv. Then: w = Xinv * tnew."

Unless Xinv is used for something else in the algorithm, then I'd say
that the author really meant "use PLS to find w."

You might get a more authoritative answer if you cited the source of
the algorithm.
.

Follow-Ups:
- Re: pseudoinverse matrix
  - From: Tim De Meyer

References:
- pseudoinverse matrix
  - From: Tim De Meyer

Prev by Date: Home-made program for FET
Next by Date: Re: Approxtn to LH Tail of NC F-Distbtn
Previous by thread: pseudoinverse matrix
Next by thread: Re: pseudoinverse matrix
Index(es):
- Date
- Thread

Relevant Pages

"Dolphin Encrypt" cryptanalysis?
... called "Dolphin Encrypt," apparently originally authored by one Peter ... Meyer. ... on the algorithm, but what I found was vague and incomplete. ...
(sci.crypt)
Re: pseudoinverse matrix
... tnew. ... This is done by a PLS estimation giving a ... Unless Xinv is used for something else in the ... the algorithm. ...
(sci.stat.math)