[topicmapmail] Merging of Distributed Topic Maps based on the Subject Identity Measure (SIM) Approach

[Bernard Vatant]

Hello Lutz

Really interesting stuff!

Just a general remark, before maybe more technical ones come from others : by an amazing
coincidence, two weeks ago, I launched a blog [1] that I want to focus on the issue of
subject identification. The last post two days ago [2] was about identification as
experimental protocol ... and certainly SIM is really a step forward in this direction.

[1] http://universimmedia.blogspot.com/
Sort of brainstorming place for this kind of issues. Jack Park has already joined as a
collaborator.
Anyone interested to jump in, through comments and posts, is welcome.

[2] http://universimmedia.blogspot.com/2004/08/identification-as-experimental.html


Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
bernard.vatant@mondeca.com

Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
bernard.vatant@mondeca.com

-----Message d'origine-----
De : topicmapmail-admin@infoloom.com [mailto:topicmapmail-admin@infoloom.com]De la part de
Dipl.-Wirtsch.-Inf. Lutz Maicher [Universität Leipzig]
Envoyé : mardi 31 août 2004 15:50
À : topicmapmail@infoloom.com
Objet : [topicmapmail] Merging of Distributed Topic Maps based on the Subject Identity
Measure (SIM) Approach


Dear all,

back from summer holiday my colleague and I finished our paper "Merging of Distributed
Topic Maps based on the Subject Identity Measure (SIM) Approach" (see
http://www.lutzmaicher.de/PUBLIKATIONEN/%5Bmaic04b%5D.pdf or
http://www.informatik.uni-leipzig.de/~maicher/forschung2.html) for LIT 2004.

We adress with this paper the question how distributed topic maps can be merged if the
subjects of their topics aren't described by a centralised vocabulary (PSIs). We developed
a (lightweighted) approach wich yields for our example set very good results.

For further research we need general feedback from the community. Do we need such methods?
Where shortcomings of our approach are seen? etc.

In addition we are searching answers to the following three questions:
1. Which other methods (for example from schema matching) will be fruitful to use in the
field of automatic topic map merging?
2. Which topic map metrics exist? We need objective criteria for topic maps to evaluate
the quality of different methods (correlation between such metrics and the yielded
quality).
3. Where we can find other example sets? We need additional topic maps for evaluation
purposes.

We are highly interested in your comments,
greatings from gray Leipzig
Lutz Maicher

_________________________________________________________________________________
Dipl.-Wirtsch.-Inf. Lutz Maicher
Graduiertenkolleg Wissensrepräsentation | Universität Leipzig
Abteilung Automatische Sprachverarbeitung | Institut für Informatik | Augustusplatz 10-11
| 04109 Leipzig

fon 0341 97 32 303 | mail maicher@informatik.uni-leipzig.de
http://www.informatik.uni-leipzig.de/~maicher/