Empirical Study on Method-level Refactoring Using Machine Learning

Because of the importance of software refactoring for software code quality and stability, this research primarily emphasizes whether refactoring can be vital to identify probable software components for future refactoring. Modularity, reusability, modifiability, maintainability, and service-oriented development may all be improved with refactoring. This fact encourages academics to develop a new and improved machine learning paradigm for restructuring OO software. We have made a multi-purpose optimization effort to assess the OOP-based software systems or components refactoring in this work. This research intends to exploit and optimize OOP software metrics to examine code quality by performing refactoring. Our objective is to develop a highly resilient and efficient ensemble computing model for refactoring prediction at the method level into a machine learning framework using software metrics as features. The focus is on applying enhanced state-of-art data acquisition, data preprocessing, data imbalance resilient re-sampling, feature extraction, and selection, followed by improved ensemble-based classification. This work will also focus on the types of project work for different kinds of classification.