Two-phase Feature Optimization for Fast and Accurate Software Defect Prediction
International Journal of Software Engineering and Knowledge Engineering · 05.08.2026 · 1-22
DOI: 10.1142/s0218194026500592 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Software Engineering and Knowledge Engineering, online erschienen am 05.08.2026. Verfasst von Shenghan Gao und 1 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Gefördert von: Natural Science Foundation of Beijing Municipality, Education and Teaching Research Fund of the Beijing University of Technology.
Abstract (Originalsprache)
In software defect prediction, improving model efficiency and training speed is as important as maintaining accuracy. Traditional machine learning and deep learning methods often rely on a large number of features, which can lead to excessive complexity and slower training times without providing significant gains in predictive power. These challenges are particularly prominent in complex software architectures, where understanding feature dependencies and interactions is critical for effective defect prediction. Reducing model complexity and enhancing training efficiency through the generation and analysis of feature interactions is therefore essential for optimizing performance. In this paper, we propose a two-phase training approach integrated with Synthetic Minority Over-sampling Technique (SMOTE) to address the inherent class imbalance in defect data. In the first phase, Mutual Information (MI) and Lasso regularization are applied to reduce feature redundancy, filtering out irrelevant features and retaining only those most predictive of defects. This streamlined feature extraction aggressively shrinks the feature space, directly reducing the overall model complexity. In the second phase, we focus on exploring multi-feature interactions by generating interaction terms using polynomial feature expansion. Subsequently, Partial Dependence Plots (PDP) are employed strictly as a post-hoc interpretability tool to visualize and validate the impact of these high-order relationships, ensuring the model captures essential feature synergies without altering original code semantics. Extensive experiments on the BugHunter dataset demonstrate the practical superiority of our approach. By explicitly decoupling offline feature preparation from online model fitting, we reveal that our dimensional reduction leads to a massive 83.7% decrease in pure model fitting time for baseline classifiers like Logistic Regression. More importantly, comparative evaluations against published state-of-the-art (SOTA) methods, including mRMR and AutoSpearman, confirm that our method achieves an optimal engineering trade-off between computational efficiency and predictive robustness. When applied to complex architectures such as Neural Networks, our selected feature subset maintains a highly competitive F1-score of 0.478 and a Matthews Correlation Coefficient (MCC) of 0.342, sacrificing less than 3% in accuracy compared to the 95-feature baseline to achieve immense speedups. Furthermore, perturbation-based consistency evaluations demonstrate a 97% stability rate across iterations, proving that the extracted interactions represent robust architectural patterns rather than statistical anomalies. Ultimately, this method simplifies the model while preserving robust performance, making it exceptionally suitable for rapid, resource-constrained continuous integration (CI/CD) environments.
Förderung
- Natural Science Foundation of Beijing Municipality
- Education and Teaching Research Fund of the Beijing University of Technology
Zitieren
APA
Gao, S., & Wang, W. (2026). Two-phase Feature Optimization for Fast and Accurate Software Defect Prediction. International Journal of Software Engineering and Knowledge Engineering, 1-22. https://doi.org/10.1142/s0218194026500592
Vancouver
Gao S, Wang W. Two-phase Feature Optimization for Fast and Accurate Software Defect Prediction. International Journal of Software Engineering and Knowledge Engineering. 2026:1-22. doi:10.1142/s0218194026500592
BibTeX
@article{Gao2026,
author = {Gao, Shenghan and Wang, Weidong},
title = {Two-phase Feature Optimization for Fast and Accurate Software Defect Prediction},
journal = {International Journal of Software Engineering and Knowledge Engineering},
year = {2026},
pages = {1--22},
doi = {10.1142/s0218194026500592}
}
EUVIMED