>[I posted this before but got no response, so I'm trying again.]
>
>Can anyone tell me if belief-function theories such as Dempster-Shafer are
>capable of handling continuous domains? Shafer's book only discusses finite
>domains, and it's not at all obvious to me how one would generalize it to a
>continuous domain.
You should read
1. Smets Ph. (1994) Belief induced by the Knowledge of some Probabilities.
Uncertainty in AI'94, Heckerman D., Poole D., & Lopez de Mantaras R. (eds),
Morgan Kaufmann, San Mateo, (1994) 523-530.
The text is on my web site, iunfortunately not the image (thanks to
Microsoft). Get the proceedings. if you can't I'll send a paper copy upon
request. You will see many equation ofr bf on the [0,1] intervals. I can go
the larger spaces, but don't be too general please)
2. Look at papers by Shafer. In his thesis (+/-73) he discussres it, and so
does he is a later paper mid 80's (As far as I remember, there is paper on
it in IJAR or Int J. Intel Systems.
3. See Tom Strat's paper, Proceeding AAAI-84, Continuoius bf for evid
reasoning.
4. The picture (triangle) you'll find in Smets 94 and in Stratt 84 papers
was already in my 1978 thesis (in french!!!!)
Tell me what you want to do, because I still have several unpublished
results in my files.
Good luck. Look at my web site, you 'll find many papers on bf. Look first
at TBM-Hdbk
Smets Ph. (1998) The Transferable Belief Model for Quantified Belief
Representation.
Handbook of defeasible reasoning and uncertainty management systems. Gabbay
D. M. and Smets Ph. Eds. Vol. 1, Kluwer, Doordrecht, 1998, 267-301.
it is an up to date survey of the TBM (i.e., bel fct seen as NOT lower proba)
Yours
Philippe
********************************************
Philippe Smets
email: [EMAIL PROTECTED]
web: http://iridia.ulb.ac.be/~psmets
IRIDIA-CP 194/6, Universite Libre de Bruxelles
50 av. Roosevelt,
1050 Bruxelles, Belgium.
tel 32 2 650 27 29 secretary,
32 2 344 82 96 private (where I am usually)
fax: 32 2 650 27 15
GSM: 32 495 50 10 72
********************************************
