Numerical Data Clustering Ontology Approach

Abstract

Clustering algorithm tasks are used to group given objects defined by a set of numerical properties in such a way that the objects within a group are more similar than the objects in different groups. All clustering algorithms have common parameters the choice of which characterizes the effectiveness of clustering. The most important parameters characterizing clustering are: metrics, number of clusters and cluster validity criteria. In classic clustering algorithms semantic knowledge is ignored. This creates difficulties in interpreting the results of clustering. At present, the use of ontology opportunities is developing very rapidly, that provide an explicit model for structuring concepts, together with their interrelationship, which allows you to gain knowledge of a particular data model. According to the previously obtained results of clustering study, the author will make an attempt to create ontology-based concept from numerical data using similarity measures, cluster numbers, cluster validity and others characteristic features. To scientific novelty should be attributed the combination of approaches of classical data analysis and ontological approach to their structuring, that increases the efficiency of their use in engineering practice.