| Title: | Mathematics at DEC |
| Moderator: | RUSURE::EDP |
| Created: | Mon Feb 03 1986 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2083 |
| Total number of notes: | 14613 |
Calling all stats people. I need some help with statistical inference.
Given a set of faults in a circuit, and a randomly selected subset.
An algorithm is tested on the subset, and is given a grade
on its detection of faults in the subset.
Assuming that faults are randomly distributed, what kind of statement
can be made about the coverage of the entire circuit, and to what
accuracy??
I understand that this is an application of an old tried-and-true
statistical technique. If anyone could convey the details, along
with references for further reading, I would be most grateful.
David
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 738.1 | Avoid complicating the problem | RDVAX::PERRONE | Fri Jul 24 1987 10:01 | 12 | |
Unless I've missed something, it seems to me that you don't necessarily
need any fancy statistical techniques to solve your problem.
Take two candidate algorithms and run them on many subsets of many
circuits. Rate them according to which detects more faults.
If you want to go the statistical route, get a good stats. book
and read up on hypothesis testing.
BTW, you should probably give your algorithms several ratings, one
for each type of potential fault; since it is unlikely (?) that
an algorithm detects all faults with the same probability.
| |||||