Long Method Detection Using Graph Convolutional Networks

Long Method is a code smell that frequently happens in software development, which refers to the complex method with multiple functions. Detecting and refactoring such problems has been popular topic refactoring, many detection approaches have proposed. In past years, based on metrics or rules leading way long detection. However, approach deep learning also attracted extensive attention recent studies. this paper, we propose graph-based detect Method. The key point of our extended PDG (Program Dependency Graph) into Directed-Heterogeneous Graph as input graph used GCN (Graph Convolutional Network) build neural network for Moreover, get substantial data samples task, novel semi-automatic generate large number samples. Finally, prove validity approach, compared existing five groups datasets manually reviewed. evaluation result shows achieved good performance