A Refactoring Advisor for Enhanced Cloned Software: Based on Several Machine Learning Techniques

Software development is time-consuming and frequently monotonous. The risk of code getting copied and pasted increases as the number of developers working on a project rises. When a program contains numerous duplicates, the quality of the code can be raised by giving developers instructions on how to rework the clone and what needs to be refactored. Code that needs to be refactored can be found by looking at clones with an automatic refactoring advisor. The advisors may be incorporated into current IDEs. Here, we describe a cutting-edge learning technique that automatically extracts properties from discovered code clones and trains models to give developers advice on how to eliminate code duplication. We demonstrate that our method outperforms earlier methods in terms of both the accuracy of the provided refactoring recommendations and the ability to automatically extract the proper parameters for performing refactoring on code clones. In contrast to prior approaches’ Class-based approach, we created a model to distinguish between refactored and anonymous copies. We demonstrate how the learned model can be used to rate each cloned piece of code in a codebase and evaluate whether it needs to be refactored. We describe a novel method for turning anomalous refactoring clone types into unknown clone set participants, 1 which is a more reliable solution than prior work that employed thresholds of similarity to identify refactoring clones and offers a thorough analysis.