I think I can NOT agree with this one:
>(1) from information theory: a reduction in uncertainty. This has no other
>associated properties.
I work with Bayesian Statistics, where the main operation is of the form
p_n(\theta) \propto p_{0}(\theta) L(\theta | X) ,
where:
- \theta is the (vector) parameter of the statistical model,
- p_0 and p_n are the prior and posterior distributions
of the parameter \theta, and
- L(\theta | X) is the likelihood function of \theta,
given the observed data X=[x_1,x_2,...x_n]
Now: It is possible that the observation X is such that
the uncertainty* of p_n is greater than that of p_0
(* entropy, variance, or whatever you want)
Still, the information in X is statistically relevant!
-- Julio Stern
>From: "Stanley N. Salthe" <ssalthe@binghamton.edu>
>To: fis@listas.unizar.es
>Subject: Re: [Fis] Re: What is the definition of information ?
>Date: Mon, 29 Aug 2005 16:32:34 -0500
>
>re Sayed's:
>
> > I am searching for a critique on the definition of the term:
> >"Information". One definition, I am using is : Information
is a thing
> >which has only the property of conveying meaning. Thus a newspaper is
>not
> >information since it also has the property of being made up of paper,
it
> >has certain dimensions, it has print ink and so on. It carries
> >information. Shannon's theory basis its definition of information on
> >probability theory. For a thing to be more probable compared to
another
> >it has to be different. This differentness means
"different". For one
> >unit to be recognizable from another it has to have some
differentness.
> >Can I get a critque of this? Also citations and directions to
>bibliography
> >will be welcomed. Syed Ali
>
>While I can't comment on formal definitions, I can make some possibly
>useful comments:
>
>I use simultaneously two definitions:
>(1) from information theory: a reduction in uncertainty. This has no other
>associated properties.
>(2) functionally: any constraint on entropy production. The latter is a
>handy general way to refer to any material activity whatever. Again, while
>it refers to some kind of material configurations, and is always the
result
>of historical contingency, as a definition no particular other properties
>are referred to (but are implied).
>
>You definition involving "meaning" is really a semiotically
inspired one.
>My (2) above can be associated with meanings.
>
>STAN
>
>
>
>_______________________________________________
>fis mailing list
>fis@listas.unizar.es
>http://webmail.unizar.es/mailman/listinfo/fis
_______________________________________________
fis mailing list
fis@listas.unizar.es
http://webmail.unizar.es/mailman/listinfo/fis
Received on Tue Aug 30 17:03:15 2005
