[Top] [All Lists]

Re: [ontolog-forum] Merging Service Ontologies to form ApplicationOntolo

To: "'[ontolog-forum] '" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Rich Cooper" <rich@xxxxxxxxxxxxxxxxxxxxxx>
Date: Sat, 13 Feb 2010 11:04:02 -0800
Message-id: <20100213190410.F0E62138D24@xxxxxxxxxxxxxxxxx>



Re Doug Foxvog et al’s paper at

http://www.vitvar.com/doc/WI2007-WangVHF.pdf ,


The definition of an ontology, which the authors adapted from Karlsruhe (see the paper) :


An ontology with data types is a structure


 O := (C;·C; T;R; ¾R;·R; A; I), consisting of

a set of concepts C                               

aligned in a hierarchy·C,

a set of relations R

with ·R,

the signature ¾R : R ! C £ C,

a set of datatypes T

with type transition functions,


a set of attributes A                           

and instances I,


and the signature ¾A : A ! C £ T.


For a relation r 2 R, we define its domain and its range by

       dom(r) := ¼1(¾R(r))

and range(r) := ¼2(¾R(r)).


Then they added metadata to describe types of relations in terms of symmetric, reflexive, etc.  


This definition is extended by four concept relations with

fuzzy weights, see Section 5.2.


and then they reach a conclusion that defining “similarity” between two ontologies is still too tough for the paper to tackle …


For measuring of ontology similarity sim(O1;O2) 2 [0; 1],

we consider concepts, concept relations, types and attributes. The concept instance similarity is out of this paper.


This could give us a starting point about how to interface two ontologies, if anyone is interested in this  topic.





Rich Cooper


Rich AT EnglishLogicKernel DOT com

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Rich Cooper
Sent: Saturday, February 13, 2010 10:23 AM
To: '[ontolog-forum] '
Subject: Re: [ontolog-forum] Merging Service Ontologies to form ApplicationOntologies


That paper below was written by Doug Foxvog (of this list) among others; it has an interesting overview algorithm section around the third page or so:


6. Building Application Ontologies

In order to build meaningful a (AO), a knowledge base,

a thesaurus or dictionary is necessary. WordNet is a quite

mature conceptual thesaurus with hierarchally structured

organisation and multiple semantic relations between concepts.

It is a good choice also due to its broad concept structure

of synonym sets.


6.1 Process of Building (AO)

We assume there are k semantic services from the same

application domain as the inputs of (AO) builder. As depicted

in Fig. 4, the building process is iterative including

following steps.


1. Initialize AO0 as an empty set, ;.

2. Import a SOi; 1 · i · k; re-organise it into the ontology

            [ my note: 1 <= i <= k  ]


structure defined by Definition 1.


3. For each concept of SOi, extract its relative concept

information from WordNet by the WCEA algorithm of

Section 6.2; add the retrieved information to SOi, then

get a new SO0i with more concept relations.

4. Merge SO0i and AOi¡1 using the WOMA algorithm

defined in Section 6.3 to obtain a new AOi.

5. Manage AOi, including cleaning concept conflicts,

updating and storing (see Section 6.4).


6. Repeat steps 2 to 5, until i = k, then stop.

The output of this process is an application ontology AOk,

which can be referred to by other service providers or users

to create new applications as depicted in Fig. 4.




This algorithm seems like a very useful place to start from if the list wants to consider how ontologies might be merged in daily practice.  





Rich Cooper


Rich AT EnglishLogicKernel DOT com


Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)

<Prev in Thread] Current Thread [Next in Thread>