26375
2014-07-17 19:46:51 +0000
Idea: pseudoinverse() method
2014-07-18 20:11:28 +0000
1
1
1
Unclassified
FXTF
Geometry
unspecified
PC
All
RESOLVED
LATER
P2
normal
---
1
a.cavalcanti
zcorpan
dschulze
mike
sideshowbarker+geometry
oldest_to_newest
109071
0
a.cavalcanti
2014-07-17 19:46:51 +0000
It is known fact that not all matrix can have an inverse calculated.
But a pseudoinverse (a.k.a. Moore-Penrose inverse) is always defined and has a set of interesting attributes and can be calculated through SVD (Single Value Decomposition).
What if we have a ::pseudoinverse() method for Matrix object?
109088
1
dschulze
2014-07-18 19:21:18 +0000
(In reply to Adenilson Cavalcanti from comment #0)
> It is known fact that not all matrix can have an inverse calculated.
>
> But a pseudoinverse (a.k.a. Moore-Penrose inverse) is always defined and has
> a set of interesting attributes and can be calculated through SVD (Single
> Value Decomposition).
>
> What if we have a ::pseudoinverse() method for Matrix object?
That is a really interesting idea. For the first level I would like to keep a narrow scope and suggest deferring the idea for Level 2. Is that ok with you?
109094
2
a.cavalcanti
2014-07-18 19:54:38 +0000
Level2 it is! :-)