Re: [UAI] unsupervised learning (and Bayesian methods).

From: David L Dowe (dld@cs.monash.edu.au)
Date: Thu May 11 2000 - 07:56:37 PDT

  • Next message: Alta de Waal: "[UAI] Non-Parametric Bayes"

       Dear H.Mallinson and UAI people,

          Have a look at my clustering page,
    http://www.csse.monash.edu.au/~dld/cluster.html

    I would recommend the algorithm, Snob, that I have been involved in:
    http://www.csse.monash.edu.au/~dld/Snob.html , which is good with noisy
    data (and is Bayesian). But, of course, have a look at the other offerings on
    the http://www.csse.monash.edu.au/~dld/cluster.html page, some of which are
    Bayesian.

    The term "non-parametric Bayesian" seems to mean different things to different
    people. Some people would say 'No' to your question by definition, and others
    would say 'Yes'. I shall play it safe here and now, and I shall pass.

    Regards. - David Dowe.

    Dr. David Dowe, School of Computer Science and Software Eng.,
    Monash University, Clayton, Victoria 3168, Australia dld@cs.monash.edu.au
    Tel:+61 3 9905-5776 Fax:+61 3 9905-5146 http://www.csse.monash.edu.au/~dld/
    http://www.csse.monash.edu.au/~dld/Snob.html
    http://www.csse.monash.edu.au/~dld/cluster.html
    And, at http://www.thehungersite.com/ , you can help feed the world.

    > From owner-uai@ghost.CS.ORST.EDU Thu May 11 01:53:29 2000
    > Subject: [UAI] unsupervised learning.
    > To: uai@cs.orst.edu
    > Content-transfer-encoding: 7BIT
    >
    > I wish to ask about unsupervised learning.
    > Can anyone recommend a resource for algorithms effective on
    > low dimensional but noisy data?
    > Fuzzy cmeans and EM have been ineffective. There is a suspicion that
    > the Gaussian assumption is weak.
    >
    > Is there a reference for material on bayesian techniques in this area?
    > Is there such a thing as a non-parametric bayesian technique?
    >
    > Apologies if this is the wrong place to post such a request.



    This archive was generated by hypermail 2b29 : Thu May 11 2000 - 08:02:36 PDT