Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a matrix that I obtained from theoretical computation and I have another matrix which I obtained by actual data collection. How do I compare the two matrices? How do I state that one matrix is significantly different from the other?

I have looked into:

  • Frobenius Norm but I don't know what should be my threshold for comparison.
  • Looking at max and min deviation of one matrix from the other. But again, what's a good threshold?

I thought that to say one matrix varies significantly from the other, I would need to quantify the variance of those terms and compare how the deviation compares to the variance (or SD) but sadly, I don't have data for that and neither can I theoretically estimate any variances.

I am kind of lost in what would be a good metric.

share|cite|improve this question
This depends entirely on your application. Different metrics will be appropriate for different purposes. – Chris Eagle Nov 29 '12 at 17:56
Mathematical Statistics. – Inquest Nov 29 '12 at 18:01
How was the matrix constructed and what do its entries represent? – Fly by Night Nov 29 '12 at 18:59
@FlybyNight. The theoretical matrix is constructed using regression and the obtained matrix is well, obtained. – Inquest Nov 29 '12 at 19:54
In that case, the metric will be constructed using metric theory and it's details will be found, somehow. – Fly by Night Nov 29 '12 at 21:27

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.