W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > May 2008

RE: Multi-layered Knowledge Representations for Healthcare (was RE: An argument for bridging information models and ontologies at the syntactic level)

From: Kashyap, Vipul <VKASHYAP1@PARTNERS.ORG>
Date: Sat, 31 May 2008 07:36:38 -0400
Message-ID: <DBA3C02EAD0DC14BBB667C345EE2D12402948578@PHSXMB20.partners.org>
To: <dan.russler@oracle.com>
Cc: "Samson Tu" <swt@stanford.edu>, <public-semweb-lifesci@w3.org>, <public-hcls-coi@w3.org>, <Elkin.Peter@MAYO.EDU>
Hi Dan,
Apologies for the delayed response. Was caught up with a bunch of things and
wasn't able to get some serious "think" time to 
respond to this.
Peter is right that the term "EAV" is a data schema implementation model, even
though it maps directly to a classic proposition model with subject, predicate,
and object of the predicate.
[VK] Agreed. We are actually referring to the basic frame-slot-value or RDF-like
subject-predicate-object triples which any information model/ontology can be
decomposed into. 

Layer 0 then would be the most abstract layer consisting purely of formal
propositions. In this layer, some propositions may express relationships between
one or two other propositions, but otherwise, no grouping of propositions
(classes) nor inheritance are characteristic of this layer.
[VK] Agreed. There are no semantics related to generalization or aggregation
expressed at this layer. However, the basic building blocks of  things like
Metaclasses and Metaproperties can be defined in this layer. Further classes can
be defined as instances and relationships like subClassOf be defined as an
instance of a metaProperty... and son... In some sense, a similar layering is
being done withi 

Peter brings up a good point about the need to deal with belief and values in
the model. After all, an ontology is really a belief system asserted by one or
more people. How does one bring belief into a model, e.g. realism, creationism,
[VK] Can you give me some examples of belief? My initial guess is belief is sort
of orthogonal to the layers, i.e., depending on the type of belief being modeled
it may appear at any of the layers. The RDF specification does provide
constructs for reification which might be used as  one mechanism to represent
belief, but it can be tricky

Regarding your note below on Layer 2...The question is whether there are finer
layers of distinction between level 1 and layer 2 (before one actually creates
instances that apply to individual patients)?
[VK]  Can you give me some examples of things that would need finer layes of
distinction between layer 1 and layer 2?
Also: I am attaching an extract of a paper on an old project I worked on called
the InfoSleuty project when I was at MCC. Some of the ideas expressed
above are presented in a figure in that extract. In that figure
Layer 0 = Frame Layer
Layer 1 = MetaModel Layer
Layer 2 = Ontology Layer
Layer 3 is not shown.
Dan, Samson, Peter and others: I was wondering if there would be interested in
starting an HCLSIG Task specially on this topic?
Look forward to comments and feedback and apologies for the delay.

The information transmitted in this electronic communication is intended only
for the person or entity to whom it is addressed and may contain confidential
and/or privileged material. Any review, retransmission, dissemination or other
use of or taking of any action in reliance upon this information by persons or
entities other than the intended recipient is prohibited. If you received this
information in error, please contact the Compliance HelpLine at 800-856-1983 and
properly dispose of this information.

Received on Saturday, 31 May 2008 11:37:29 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:20:34 UTC