W3C home > Mailing lists > Public > www-rdf-interest@w3.org > December 1999

Certainty Factors

From: Danny Ayers <Danny.Ayers@highpeak.ac.uk>
Date: Fri, 10 Dec 1999 10:06:32 +0000
Message-ID: <3850D0A8.D2DF3939@highpeak.ac.uk>
To: "www-rdf-interest@w3.org" <www-rdf-interest@w3.org>
Hi,
I'm just wondering what angle the proposed RDF spec takes for dealing
with uncertainty, and the 'truth' of a statement. As an extreme case
let's have the assertion in one place (abstracting it a bit)

Object : Freddy's cat
Attribute : Colour
Value : Red

 and elsewhere we have the assertion
Object : Freddy's cat
Attribute : Colour
Value : Green

As I see it, there are essentially two approaches available - 1. there
is no need for implicit definition of certainty within RDF; 2. certainty
can be expressed in a RDF triple.

With 1, it could be left to a reasoning engine to determine which of
these statements is more likely to be true - following a chain of
authority perhaps, if the chain leads back to a document with Freddy as
the author, then we take his word on it.

With 2, we could have additional assertions along the lines of :

Object : The first statement about the colour of Freddy's cat
Attribute : Certainty
Value : 0.9

or even

Object : This document
Attribute : Certainty
Value : 0.01

Any thoughts on this?

I must confess that my knowledge of RDF is pre-pubescent, and apologies
if I've missed a relevant thread or document.

Cheers,
Danny.











 Having looked at other knowledge representation methods in the past 



-------------

Click on a button on this site to make a *free* food donation to hungry
people (paid for by sponsors) :
http://www.thehungersite.com/index.html
Received on Friday, 10 December 1999 05:07:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:42 GMT