A cosine theorem based algorithm for similarity aggregation of ontologies

Ontology matching is a basic method for resolving ontology heterogeneity. In order to achieve the finest matching result through aggregation, a couple of measuring methods are often applied to the process of ontology matching and then the results of all methods produced are aggregated. This paper proposes a new method in aggregating the similarity through analysis of existing methods. Our proposal considers two similarities to be aggregated as two vectors with a specific angle and their values as the modulus of the vectors. The result of aggregation equals to the modulus of the vector determined by the subtraction of the two previous vectors. The process of calculation is involved in the application of cosine theorem and therefore that is the reason why this method is called a Cosine Theorem Based Algorithm (CTBA) for similarities aggregation of ontologies in this paper. The results of experiments and analysis conducted with the known data suggest that CTBA presented in this paper has a comparatively good performance in aggregating similarities of ontologies.