Computer Science > Software Engineering
[Submitted on 24 Jun 2024]
Title:Cascade Generalization-based Classifiers for Software Defect Prediction
View PDFAbstract:The process of software defect prediction (SDP) involves predicting which software system modules or components pose the highest risk of being defective. The projections and discernments derived from SDP can then assist the software development team in effectively allocating its finite resources toward potentially susceptible defective modules. Because of this, SDP models need to be improved and refined continuously. Hence, this research proposes the deployment of a cascade generalization (CG) function to enhance the predictive performances of machine learning (ML)-based SDP models. The CG function extends the initial sample space by introducing new samples into the neighbourhood of the distribution function generated by the base classification algorithm, subsequently mitigating its bias. Experiments were conducted to investigate the effectiveness of CG-based Naïve Bayes (NB), Decision Tree (DT), and k-Nearest Neighbor (kNN) models on NASA software defect datasets. Based on the experimental results, the CG-based models (CG-NB, CG-DT, CG-kNN) were superior in prediction performance when compared with the baseline NB, DT, and kNN models respectively. Accordingly, the average accuracy value of CG-NB, CG-DT, and CG-kNN models increased by +11.06%, +3.91%, and +5.14%, respectively, over baseline NB, DT, and kNN models. A similar performance was observed for the area under the curve (AUC) value with CG-NB, CG-DT, and CG-kNN recording an average AUC value of +7.98%, +26%, and +24.9% improvement over the baseline NB, DT, and kNN respectively. In addition, the suggested CG-based models outperformed the Bagging and Boosting ensemble variants of the NB, DT, and kNN models as well as existing computationally diverse SDP models.
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.