Model Matching: Prediction of the Influence of UML Class Diagram Parameters During Similarity Assessment Using Artificial Neural Network

Abstract

Model matching is an approach in which two or more software systems are compared. The comparison requires the computation of similarity between models in different systems. The idea is to find the most similar model and be re-used in another context especially during the development of a new software system. Models such as UML class diagrams consist of several parameters (such as class name); these parameters are used during the similarity assessment between class diagrams. This chapter investigates the influence of different parameters used during the computation of similarity between class diagrams using artificial neural network (ANN).