AN ONTOLOGY-BASED CLUSTER ANALYSIS FRAMEWORK

Poster and Demonstration Session at the 7th International Semantic Web Conference

计算机科学;图书情报档案学

Paweł Lula ; Grażyna Paliwoda-Pękosz

Others  :  http://CEUR-WS.org/Vol-401/iswc2008pd_submission_68.pdf
PID  :  25507

来源: CEUR

【 摘 要 】

The main objectives of this paper are to discuss the various aspects of similarity calculations between objects and sets of objects in ontology-based environments and to propose a framework for cluster analysis in such an environment. The framework is based on the ontology specification of two core components: description of categories and description of objects. Similarity between objects is defined as an amalgamation function of taxonomy, relationship and attribute similarity. The different measures to calculate similarity that can be used in framework implementations are presented. The ontology-based data representation and the framework of cluster analysis can be useful in the area of Business Intelligence, e.g. clustering similar companies that profiles are described by ontology-based data.