Extension for Design Pattern Identification Using Similarity Scoring Algorithm

This paper provides studies in area of design pattern identification in existing software systems. The target was to design and prototype specific extensions of a selected method with better and more precise results. This paper summarizes separate phases of a project analysis of different design pattern representations, analysis of different methods, algorithms for identification of design patterns in software systems, selection of a concrete method for extensions, design, implementation and tests of an extended method. In this paper, the extensions like feature weighting, feature filtering and additional lexicographical analysis are described. Designed and prototyped extensions were experimentally tested on more than 3000 classes of open source systems from different software engineering areas. Based on the test results, the general conclusion for designed extensions are formulated.