This survey focuses on software defect prediction (SDP) in web applications, systematically analyzing recent advances in feature selection-based SDP methodologies. It aims to investigate the impact of different feature selection techniques, including filter, wrapper, embedding and hybrid methods, on prediction performance while summarizing current research trends and challenges.
To systematically analyze and summarize this research question, the authors collected relevant papers published between 2020 and 2024 from academic databases including Google Scholar, DBLP, ACM Digital Library, Springer Link, Web of Science and IEEE Xplore, using keywords such as Web application defect prediction (WADP), SDP, feature selection and software fault prediction.
This literature review, based on 40 core studies carefully selected from IEEE Xplore, Science Direct, ACM Digital Library and Springer Link, provides profound and valuable insights into the critical role of feature selection methods in WADP.
This study explores key research questions and highlights the popularity trends and notable advantages of filter, wrapper, embedding and hybrid methods. These findings lay a solid foundation for a comprehensive understanding of the current trends in feature – selection methods and establish a strong groundwork for future research aimed at enhancing the accuracy and efficiency of Web application defect – prediction models.
