Date: Jan 12, 2013 11:33 PM
Author: Peter Mairhofer
Subject: Pseudo inverse: small deviations, large error

Hi,

Can someone plain why I get such huge errors with only slight deviations?

>> Hf
Hf =

  1.0e+003 *

    1.6443    1.6516    1.6583
    4.8373    4.8349    4.8334
    4.6385    4.6418    4.6445
   -9.6014   -9.6084   -9.6154

>> [yl1 , yl3 , yl1 - yl3 ]
ans =

  1.0e+006 *

    0.2966    0.2972   -0.0006
    0.8705    0.8703    0.0002
    0.8352    0.8355   -0.0003
   -1.7288   -1.7295    0.0006

This is the correct result:

>> Hf \ yl1
ans =

  100.0000
   75.0000
    5.0000

And now:

>> Hf \ yl3
ans =

   56.0412
   72.5578
   51.4007

... unuseable results. I have much redundancy available, I can make the
rows of "Hf" very large but still get bad results.


Thanks,
Peter