Re: Determining Confidence Interval
From: George Kahrimanis (anakreon_at_hol.gr)
Date: 09/30/04
- Next message: Qu Zhang: "Is it possible?"
- Previous message: George Kahrimanis: "Re: A simple but confusing question"
- Next in thread: Ion Saliu: "Re: Determining Confidence Interval"
- Maybe reply: Ion Saliu: "Re: Determining Confidence Interval"
- Messages sorted by: [ date ] [ thread ]
Date: 30 Sep 2004 09:46:26 -0700
Roger, I hope I am helping if I try to break it down.
Say, you need to know check whether the probability is
0.2 +/- 0.01, and you want a method that
(I) if this is the truth, the method would not say "false"
more often than, say 3% of the time, on average (worst case)
and
(II) if this is not the truth, the method would not
say it is right, more often than, say, 1$ of the time,
on average (worst case).
However, I recommend that you calculate what you gain and what
you lose in each of these cases, and form a decision problem,
if possible.
Good Luck, ~ George Kahrimanis
- Next message: Qu Zhang: "Is it possible?"
- Previous message: George Kahrimanis: "Re: A simple but confusing question"
- Next in thread: Ion Saliu: "Re: Determining Confidence Interval"
- Maybe reply: Ion Saliu: "Re: Determining Confidence Interval"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
