Re: same entropy between two probability distribution functions.
From: Robin Morphet (rm_at_rmjm.demon.co.uk)
Date: 01/24/05
- Next message: George Cox: "Re: article on schools of foundations of mathematics"
- Previous message: George Cox: "Re: article on schools of foundations of mathematics"
- In reply to: Yan ZHANG: "Re: same entropy between two probability distribution functions."
- Next in thread: beliavsky_at_aol.com: "Re: same entropy between two probability distribution functions."
- Messages sorted by: [ date ] [ thread ]
Date: Mon, 24 Jan 2005 23:41:09 +0000
If you consider a histogram (unimodal for convenience) and then exchange
one of the central bars with one at the end of the distribution the
distribution could change markedly. However, the entropy would remain
the same. With a measure such as chi square the difference in shape
between the two distributions would register in most cases.
In message <35b9vbF4j8r48U1@individual.net>, Yan ZHANG
<buaanupt@sina.com> writes
>Thank you very much for your comments.
>
>Could you plz give more explanation why using the same entropy as the
>criteria is not a good idea?
>
-- Robin Morphet
- Next message: George Cox: "Re: article on schools of foundations of mathematics"
- Previous message: George Cox: "Re: article on schools of foundations of mathematics"
- In reply to: Yan ZHANG: "Re: same entropy between two probability distribution functions."
- Next in thread: beliavsky_at_aol.com: "Re: same entropy between two probability distribution functions."
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
