=Paper=
{{Paper
|id=Vol-500/paper-1
|storemode=property
|title=Ontology Merging by Matching a Selection of Ontologies in a Cluster Environment
|pdfUrl=https://ceur-ws.org/Vol-500/paper1.pdf
|volume=Vol-500
}}
==Ontology Merging by Matching a Selection of Ontologies in a Cluster Environment==
<pdf width="1500px">https://ceur-ws.org/Vol-500/paper1.pdf</pdf>
<pre>
Ontology Merging by Matching a
  Selection of Ontologies in a
     Cluster Environment
     PhD topic (abstract): Axel Tenschert, tenschert@hlrs.de

  High Performance Computing Center Stuttgart, University of
                        Stuttgart



Ontology matching is of high interest for a number of tasks solved in the semantic
web, especially when thinking of extending complex information structures. With
regard to this work current matching strategies will be considered in order to create
a strategy for matching ontologies with the aim to merge them. The purpose of such
a merging is to support a user (e.g. scientist, doctor, … ) with one priority ontology
which is related to his specific requirements. Hereby, the user is enabled to use the
knowledge of several ontologies by using only the one priority ontology.
Furthermore, when thinking of merging ontologies to one priority ontology, the
research field of bioinformatics is of interest because of the need to investigate large
scale data within a high number of ontologies. The benefit for such a user is the
possibility to receive required information very fast by considering only one data
source, the extended priority ontology.

There are several ontology matching strategies and merging tools available.
Nevertheless, the complexity of matching ontologies entails the problem of matching
them in a scalable way. In order to solve this problem distribution and parallelization
techniques are used to speed up the matching by executing it at same time. For this
the matching process is distributed to execute as much comparisons between
concepts at same time as possible. The concepts from the priority ontology are
matched with the concepts from the set of ontologies in parallel. It is obvious that
this methodology requires lots of computing resources to execute the large amount
of data.

To provide the required computing resources the matching of the concepts and the
merging is executed in a cluster environment. The selection of ontologies is copied on
the hard drive of the cluster. Afterwards, the matching and merging procedures are
executed as several jobs on the nodes of the cluster. When all matching concepts are
merged together in the priority ontology, the ontologies on the hard drive of the
cluster are deleted automatically expect the extended priority ontology. This priority
ontology is prepared to be used by a user, e.g. a doctor or a scientist in the field of
medicine/bioinformatics.

</pre>