Re: estimating the number of false positives
clemenr_at_wmin.ac.uk
Date: 03/02/05
- Next message: Richard Ulrich: "Re: estimating the number of false positives"
- Previous message: Dominique Vlieghe: "estimating the number of false positives"
- In reply to: Dominique Vlieghe: "estimating the number of false positives"
- Next in thread: Richard Ulrich: "Re: estimating the number of false positives"
- Messages sorted by: [ date ] [ thread ]
Date: 2 Mar 2005 05:45:10 -0800
Disclaimer: I'm no expert statistician but that rarely shuts me up.
For what it's worth, I agree with you. If the chance of an observation
being "yes" is p, and the only other possible outcome is "no", then the
probability of "no" for that observation is q = (1-p). Hence for any M
element subset of your N observations, the expected number of "no"
results would be the SUM of qi for i=1,2,3,..,M.
If you're looking at how the accuracy of a test as you vary a
threshold, and if you're not looking at them already, you might want to
google up Receiver Operating Characteristic Curves.
Cheers,
Ross-c
- Next message: Richard Ulrich: "Re: estimating the number of false positives"
- Previous message: Dominique Vlieghe: "estimating the number of false positives"
- In reply to: Dominique Vlieghe: "estimating the number of false positives"
- Next in thread: Richard Ulrich: "Re: estimating the number of false positives"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
