Beefy Boxes and Bandwidth Generously Provided by pair Networks
good chemistry is complicated,
and a little bit messy -LW

Re: Pseudoinversing matrix .. is this correct?

by tall_man (Parson)
on Aug 11, 2005 at 23:18 UTC ( #483146=note: print w/replies, xml ) Need Help??

in reply to Pseudoinversing matrix .. is this correct?

I see a couple of problems here. First, there seems to be confusion between the mxn matrix B (which you assigned to the variable $matrix) and the diagonal matrix in the decomposition (which you assigned to the variable $b).

Secondly, the diagonal result in $b is returned in the form of a single row. It must be converted to a diagonal matrix before you can work on it.

I would do something like this instead:

my $mt = transpose($matrix); my $m2 = $mt x $matrix; my $m2i = $m2->inv; my $psuedo = $m2i x $mt;

You will only need svd if the m2 matrix is singular or ill-conditioned. It will allow you to "fix" any near-zeros on the diagonal.

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://483146]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others exploiting the Monastery: (5)
As of 2016-10-22 19:53 GMT
Find Nodes?
    Voting Booth?
    How many different varieties (color, size, etc) of socks do you have in your sock drawer?

    Results (297 votes). Check out past polls.