How many runs to statistically validate a 10-DOF algorithm?

  • Context: Undergrad 
  • Thread starter Thread starter rickdatech
  • Start date Start date
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
1 reply · 2K views
rickdatech
Messages
1
Reaction score
0
I understand a bit about math, but it's been over 30 years since I've been to college. I have a question about where to find information mostly. My prolbem is that I have an algorithm that guesses the results of a process. The process has as best I can figure about 10 degrees of freedom.

1) how many times would I need to run the algorithm against the process to statistically validate the algorithm?

2) where would I find an Internet source that discusses this type of statistical validation? none of the google terms I come up with work for me.

I understand I can prove it by showing the math of how the algorithm was constructed, but that only works if you know the "why" behind the algorithm.
 
Physics news on Phys.org
1. No less than 30.
2. Look up statistical regression analysis.