From: "c." <address@hidden>
Date: 25 August 2010 19:31:01 GMT+02:00
To: address@hidden
Subject: Re: Bug-octave Digest, Vol 53, Issue 13
On 25 Aug 2010, at 19:03, address@hidden wrote:
Not sure if this is a bug.
a = [ 1 2 3]
b = [ 2 4 6]
As expected,
a/b returns 0.50
a./b returns 0.50 0.50 0.50.
Now change b(3) = 7
a/b returns 0.44928
What is this result/calculation?
this is equivalent to
a*b'/(b*b')
or
a * pinv(b)
it is a solution of:
find x that minimizes
norm(x*b -a, 2)^2
As expected,
a./b returns 0.50 0.50 .42857
c.