Re: Bayes Destroyed?

From: marc.geddes <marc.geddes.domain.name.hidden>
Date: Fri, 28 Aug 2009 22:55:27 -0700 (PDT)

On Aug 29, 5:21†am, Brent Meeker <meeke....domain.name.hidden> wrote:

>
> Look at Winbugs or R. †They compute with some pretty complex priors -
> that's what Markov chain Monte Carlo methods were invented for.
> Complex =/= uncomputable.

 Techniques such the Monte Carlo method donít scale well.
>

>
> Actually Bayesian inference gives a precise and quatitative meaning to
> † Occam's razor in selecting between models.
>
> http://quasar.as.utexas.edu/papers/ockham.pdf
>
>

The formal definitions of Occamís razor are uncomputable. Remember,
the theory of Bayesian reasoning is *itself* a scientific model, so
differences of opinion about Bayesian models will result in mutually
incompatible science. Thatís why Bayes has serious problems. (see
below for more on this point)


>
> And beliefs do not converge, even in probability - compare Islam and
> Judaism. †Why would any correct theory of degrees of belief suppose
> that finite data should remove all doubt?


So how did people come to believe things like Islam and Judaism in
the first place? (the beliefs PRIOR to collecting evidence) Bayes
canít tell you *what* to believe, it can only tell you how your
beliefs should *change* with new evidence. The fact that you are free
to believe anything to start with shows that Bayes has major
problems.

Stathis once pointed on this list that crazy people can actually still
perform axiomatic reasoning very well, and invent all sorts of
elaborate justifications, the problem is their priors, not their
reasoning; so if you try to use Bayes as the entire basis of your
logic, youíre crazy ;)



--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Everything List" group.
To post to this group, send email to everything-list.domain.name.hidden
To unsubscribe from this group, send email to everything-list+unsubscribe.domain.name.hidden
For more options, visit this group at http://groups.google.com/group/everything-list?hl=en
-~----------~----~----~----~------~----~------~--~---
Received on Fri Aug 28 2009 - 22:55:27 PDT

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:16 PST