International Conference on Computer and Knowledge Engineering

Home / 15th International Conference on Computer and Knowledge Engineering

Impact of Oversampling Methods on Imbalanced Dataset for Software Fault Prediction

Authors :

Alireza Abiri¹ Alireza Tajary² Mansoor Fateh³

1- Faculty of Computer Engineering, Shahrood University of Technology, Shahrood, Iran 2- Faculty of Computer Engineering, Shahrood University of Technology, Shahrood, Iran 3- Faculty of Computer Engineering, Shahrood University of Technology, Shahrood, Iran

Keywords :

Software Fault Prediction،Imbalanced Data،Machine Learning،GAN،BugHunter Dataset

Abstract :

In today's world, with the rapid advancement of technology and the increasing use and scale of software systems both in terms of data volume and number of users, the occurrence of software faults has become inevitable. Consequently, software fault prediction has gained significant importance for the early identification of faulty modules during the software development process. However, one of the key challenges in this domain is the class imbalance problem, where the number of faulty and non-faulty instances in software datasets is highly unequal. To address this issue, data oversampling techniques are commonly employed to balance the datasets. In this study, we investigate and compare the performance of three data oversampling methods on the BugHunter software fault dataset. The results indicate that using Generative Adversarial Networks (GANs) for data generation and oversampling is a more effective approach for addressing class imbalance, achieving better performance compared to alternative methods.