Howdy,
I am wondering if the documentation is correct for the Diagnosis page:
http://genie.sis.pitt.edu/wiki/Support_ ... tic_window
The page claims:
> The basis for ranking will be calculated for the table according to the
> following equation:
>
> E(F1) = P(F1|T1) + alpha * C1(T1).alpha
However, the actual numbers reported back look like the mutual information between the fault and the test. Would it be better to write
the value of the test T1 in isolating fault F1 in terms of mutual information?
> V(T1) = I(T1; F1) = H(F1)-H(F1|T1) + alpha * C(T1)
The attached file shows a simple example. The numerical values agree with the output of SMILE.
Diagnosis Documentation Correction
-
- Posts: 5
- Joined: Wed Sep 23, 2009 2:00 am
Diagnosis Documentation Correction
- Attachments
-
- bayesianUpdateAndInformationGain.xls
- (24 KiB) Downloaded 369 times
-
- Site Admin
- Posts: 430
- Joined: Tue Dec 11, 2007 4:24 pm
Cost of diagnosis
Thanks! It looks like you are right. The documentation was a student project and may contain errors here and there. I have just made a correction to this section, although this was a quick fix -- I hope I haven't overlooked something. If you notice a problem in the future, please feel free to edit the documentation -- we have moved the docs to the Wiki format so that everybody can make a change. Thanks again for pointing this out to us!
Marek
Marek