Determining semantic similarity of it systems based on the comparison of their graphical data models

Abstract

Modelling is the basis for research and development of IT systems. Graphical models and graphic representations of models originally built in non-graphic languages and formalisms are often used. In modelling IT systems a need exists for comparing graphical models which can represent different variations of the same or similar modelled content or graphical models which, with certain revisions, could be applied in various domains. Graphical model in the latter case first needs to be translated into another form of predicate expressions or formal languages of modelled content representation. The lack of translation of the model for comparison is a time-consuming venture and may result in the loss of modelled relations due to differences in the "language" and representation symbols.
The goal of the paper is to explore and propose methods and procedures for determining similarities of IT systems based on the comparison of their graphical data models. The procedure of determining the similarities of graphical data models of the same type shall at the end of my research be based on semantic and structural similarity of models. In this article procedures for determining semantic similarity and their application is discussed together with examples and the procedure for determining structural similarity is proposed in roughly as it is in the finishing state of current research activities.