Improving the Accuracy of Ontology Alignment through Ensemble Fuzzy Clustering

Abstract

Abstract. Automatic ontology alignment tools perform matching between the concepts of two ontologies and provide a similarity measure for each pair of aligned concepts. However, none of the existing tools are perfect and multiple alignment tools produce varying similarity measures for a certain alignment. Also, the similarity measures provided by an alignment may not be helpful enough for indicating the degree of reliability. While using a random alignment tool we noticed that some quality alignments are given medium or even low similarity measures, and that causes the user ignoring those alignments. In this study we have proposed an ensemble model of ontology alignment that aggregates multiple alignment tools with the help of Fuzzy C Means clustering and Type 2 Fuzzy Membership Functions. We have shown that our approach helps the user to choose the best alignment results which has not been obtained by any other alignment tools we experimented with.

Using Fuzzy Ontology to Improve Similarity Assessment: Method and Evaluation

“…In this way, common features tend to increase similarity and non‐common ones tend to diminish it. This approach is originated from Tversky model and has been applied in molecular biology, adaptive e‐learning and ontology merging and alignment Information‐based approaches calculates statistical specification of concepts based on a corpus or other data sources .…”