Combining Structural Features and Software Metrics for Automated Design Pattern Detection

Design patterns provide reusable solutions to recurring software design problems. Since their introduction by the Gang of Four in 1994, identifying these patterns in object-oriented programs has remained a challenging task, particularly with regard to leveraging software metrics. This paper proposes a two-phase detection approach that combines structural features with software metrics to identify design patterns. The approach was evaluated on eight subject systems, achieving a precision of 88%. The experimental results demonstrate that the hybrid use of structural analysis and software metrics is effective in detecting design pattern instances.