A fuzzy matching approach for design pattern mining

Abstract

Mining design patterns from source code is significant for improving the intelligibility and maintainability of software. In this paper, we present a new design pattern matching method based on fuzzy, in which matrix model is used for describing both design pattern and source code, and design pattern's static and dynamic information is defined as fuzzy attribute value for measuring the matching degree. Experiments on three open-source projects demonstrate the accuracy and efficiency of the proposed methodology.