Re: What is sufficient statistics?
From: Nicolas Roth (nicolas.w.roth_at_bluewin.ch)
Date: 11/15/04
- Next message: Richard Ulrich: "Re: I really need help in my assignments! PLEASE kindly help!"
- Previous message: Bob Ehrlich: "Journal ref for applied Bayesian Analysis"
- In reply to: lucy: "What is sufficient statistics?"
- Messages sorted by: [ date ] [ thread ]
Date: Mon, 15 Nov 2004 19:05:10 +0100
A sufficient statistic is a statistic that, once you have it, will allow you
to make all the inference you want without needing anything else. For the
normal distribution, we can say that once we know mu and sigma, we do not
need anymore the data, because those two statistics are "summarizing"
everything about the data.
I would suggest the following book:
Casella, Berger: Statistical Inference, ed: Duxbury
"lucy" <losemind@yahoo.com> a écrit dans le message de
news:cn65cq$7fe$1@news.Stanford.EDU...
> Can anybody point me to some references on what is sufficient statistics?
I
> heard this word today, but not sure what does it mean? and what is its
> implication?
>
> Any ideas?
>
> Thanks a lot!
>
>
- Next message: Richard Ulrich: "Re: I really need help in my assignments! PLEASE kindly help!"
- Previous message: Bob Ehrlich: "Journal ref for applied Bayesian Analysis"
- In reply to: lucy: "What is sufficient statistics?"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|