Cross-project Software Defect Prediction Using Divergence-based Domain Adaptation Approach

Abstract

Cross-project Defect Prediction (CPDP) aims to predict defects in software with no or limited training data by leveraging labeled data from similar projects. However, a cross-project learning task is challenging due to the domain shift caused by differences in feature distributions across domains. Adversarial Domain Adaptation (ADA) techniques have recently shown promise in addressing these distribution differences to improve cross-project learning performance. In this work, we propose MDD-DA-CPDP, a Cross-project Software Defect Prediction approach based on the Maximum Density Divergence Adversarial Domain Adaptation method. The Maximum Density Divergence (MDD) method minimizes inter-domain differences in data distribution while maximizing intra-domain differences. As a result, the data distribution difference between two similar projects can be effectively mitigated. The proposed MDD-DA-CPDP approach first prepares source and target-domain features and then uses them to minimize the distribution gap between projects while maximizing intra-domain differences within each domain. The MDD-DA-CPDP approach is validated through experiments on datasets from the AEEEM and PROMISE repositories using various performance metrics. Statistical tests are also conducted to assess its performance significance. The results show that the proposed approach achieved significant improvements in CPDP and consistently outperformed state-of-the-art approaches by 1.3% to 8% across different evaluation metrics.

Keywords

Domain adaptation / software defect prediction / maximum density divergence / cross-project prediction

References

[1] Bai J, Jia J, Capretz L FA three-stage transfer learning framework for multi-source cross-project software defect prediction. Information and Software Technology, 2022, 150: 106985.

[2] Bal P R, Kumar SCross project defect prediction using dropout regularized deep learning and unique matched metrics. ACM Transactions on Management Information Systems, 2025, 16(3): 1-32.

[3] Bennin K E, Tahir A, MacDonell S G, Börstler JAn empirical study on the effectiveness of data resampling approaches for cross-project software defect prediction. IET Software, 2022, 16(2): 185-199.

[4] Bhat N A, Farooq S UAn empirical evaluation of defect prediction approaches in within-project and cross-project context: Ahamad and Umar. Software Quality Journal, 2023, 31(3): 917-946.

[5] Briand L C, Melo W L, Wust JAssessing the applicability of fault-proneness models across object-oriented software projects. IEEE Transactions on Software Engineering, 2002, 28(7): 706-720.

[6] Chawla N V, Bowyer K W, Hall L O, Kegelmeyer W PSMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 2002, 16321-357.

[7] Chen D, Chen X, Li H, Xie J, Mu YDeepCPDP: Deep learning based cross-project defect prediction. IEEE Access, 2019, 7: 184832-184848.

[8] Du X, Zhou Z, Yin B, Xiao GCross-project bug type prediction based on transfer learning. Software Quality Journal, 2020, 28(1): 39-57.

[9] D’Ambros M, Lanza M, Robbes REvaluating defect prediction approaches: A benchmark and an extensive comparison. Empirical Software Engineering, 2012, 17(4): 531-577.

[10] Elish M O, Elish KAn empirical comparison of resampling ensemble methods of deep learning neural networks for cross-project software defect prediction. International Journal of Intelligent Engineering & Systems, 2021

[11] Fang Y, Yap P, Lin W, Zhu H, Liu MSource-free unsupervised domain adaptation: A survey. Neural Networks, 2024, 174: 106230.

[12] Farahani A, Voghoei S, Rasheed K, Arabnia H RA brief review of domain adaptation. Advances in Data Science and Information Engineering: Proceedings from IC-DATA 2020 and IKE 2020, 2021877-894.

[13] Fritz C O, Morris P E, Richler J JEffect size estimates: Current use, calculations, and interpretation. Journal of Experimental Psychology: General, 2012, 141(1): 2.

[14] Gangwar A K, Kumar SConcept drift in software defect prediction: A method for detecting and handling the drift. ACM Transactions on Internet Technology, 2023, 23(2):1-28.

[15] Giray G, Bennin K E, Köksal Ö, Babur Ö, Tekinerdogan BOn the use of deep learning in software defect prediction. Journal of Systems and Software, 2023, 195: 111537.

[16] Gong L, Jiang S, Bo L, Jiang L, Qian JA novel class-imbalance learning approach for both within-project and cross-project defect prediction. IEEE Transactions on Reliability, 2019, 69(1): 40-54.

[17] Gong L, Jiang S, Bo L, Jiang L, Qian JA novel class-imbalance learning approach for both within-project and cross-project defect prediction. IEEE Transactions on Reliability, 2019, 69(1): 40-54.

[18] Goodfellow I J, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio YGenerative adversarial nets. Advances in Neural Information Processing Systems (NIPS 2014), 2014December 8–13, 2014

[19] Hosseini S, Turhan B, Mäntylä MSearch based training data selection for cross project defect prediction. Proceedings of the the 12th International Conference on Predictive Models and Data Analytics in Software Engineering (PROMISE 2016), 2016September 9, 2016

[20] Hsu C, Chen C, Huang SAdaptive adversarial contrastive learning for cross-domain recommendation. ACM Transactions on Knowledge Discovery from Data, 2023, 18(3): 1-34.

[21] Huang J, Gretton A, Borgwardt K, Schölkopf B, Smola ACorrecting sample selection bias by unlabeled data. Advances in Neural Information Processing Systems (NIPS 2006), 2006December 4–9, 2006

[22] Javed K, Shengbing R, Asim M, Wani M ACross-project defect prediction based on domain adaptation and LSTM optimization. Algorithms, 2024, 17(5): 175.

[23] Jiang W, Qiu S, Liang T, Zhang FCross-project clone consistent-defect prediction via transfer-learning method. Information Sciences, 2023, 635138-150.

[24] Jin CCross-project software defect prediction based on domain adaptation learning and optimization. Expert Systems with Applications, 2021, 171114637.

[25] Kanwar S, Awasthi L K, Shrivastava VCandidate project selection in cross project defect prediction using hybrid method. Expert Systems with Applications, 2023, 218119625.

[26] Khan B, Iqbal D, Badshah SCross-project software fault prediction using data leveraging technique to improve software quality. Proceedings of the 24th International Conference on Evaluation and Assessment in Software Engineering (EASE’ 20), 2020April 14–17, 2020

[27] Khatri Y, Singh S KCross project defect prediction: A comprehensive survey with its SWOT analysis. Innovations in Systems and Software Engineering, 2022, 18(2): 263-281.

[28] Krasner HThe cost of poor software quality in the US: A 2020 report, 2021

[29] Li J, Yu H, Zhang Z, Luo X, Xie SConcept drift adaptation by exploiting drift type. ACM Transactions on Knowledge Discovery from Data, 2024, 18(4): 1-22

[30] Li J, Chen E, Ding Z, Zhu L, Lu K, Shen H TMaximum density divergence for domain adaptation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 43(11): 3918-3930.

[31] Li J, Yu Z, Du Z, Zhu L, Shen H TA comprehensive survey on source-free domain adaptation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(8): 5743-5762.

[32] Liu C, Yang D, Xia X, Yan M, Zhang XA two-phase transfer learning model for cross-project defect prediction. Information and Software Technology, 2019, 107125-136.

[33] Nevendra M, Singh PTRGNet: A deep transfer learning approach for software defect prediction. Expert Systems with Applications, 2025, 282: 127799.

[34] Omondiagbe O P, Licorish S A, MacDonell S GNegative transfer in cross project defect prediction: Effect of domain divergence. 2022 48th Euromicro Conference on Software Engineering and Advanced Applications (SEAA), 2022August 31–September 2,2022

[35] Oza P, Sindagi V A, Vishnukumar Sharmini V, Patel V MUnsupervised domain adaptation of object detectors: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 46(6): 4018-4040.

[36] Pachouly J, Ahirrao S, Kotecha K, Selvachandran G, Abraham AA systematic literature review on software defect prediction using artificial intelligence: Datasets, data validation, methods, approaches, and tools. Engineering Applications of Artificiallntelligence, 2022, 111104773.

[37] Pai A R, Joshi G, Rane SQuality and reliability studies in software defect management: A literature review. International Journal of Quality & Reliability Management, 2021, 38(10): 2007-2033.

[38] Pal S, Sillitti ACross-project defect prediction: A literature review. IEEE Access, 2022, 10118697-118717.

[39] Pan S J, Tsang IW, Kwok J T, Yang QDomain adaptation via transfer component analysis. IEEE Transactions on Neural Networks, 2010, 22(2): 199-210.

[40] Pandey S K, Tripathi A KDNNAttention: A deep neural network and attention based architecture for cross project defect number prediction. Knowledge-Based Systems, 2021, 233: 107541.

[41] Sharief F, Ijaz H, Shojafar M, Naeem M AMulti-class unbalanced data handling with concept drift in fog computing: A taxonomy, review, and future directions. ACM Computing Surveys, 2024, 57(1): 1-48.

[42] Shirabad J S, Menzies TThe PROMISE Repository of Software Engineering Databases, 2005

[43] Singhal P, Walambe R, Ramanna S, Kotecha KDomain adaptation: Challenges, methods, datasets, and applications. IEEE Access, 2023, 116973-7020.

[44] Song H, Wu G, Ma L, Pan Y, Huang Q, Jiang SAdversarial domain adaptation for cross-project defect prediction. Empirical Software Engineering, 2023, 28(5): 127.

[45] Song S, Yu H, Miao Z, Fang J, Zheng K, Ma C, Wang SMulti-spectral salient object detection by adversarial domain adaptation. Proceedings of the AAAI Conference on Artificial Intelligence (AAAI, 2020February 7–12, 2020

[46] Sun Z, Li J, Sun H, He LCFPS: Collaborative filtering based source projects selection for cross-project defect prediction. Applied Soft Computing, 2021, 99: 106940.

[47] Taheri S, Hesamian GA generalization of the Wilcoxon signed-rank test and its applications. Statistical Papers, 2013, 54(2):457-470.

[48] Tong H, Lu W, Xing W, Wang SARRAY: Adaptive triple feature-weighted transfer Naïve Bayes for cross-project defect prediction. Journal of Systems and Software, 2023, 202111721.

[49] Tzeng E, Hoffman J, Saenko K, Darrell TAdversarial discriminative domain adaptation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’ 17), 2017July 21–26, 2017

[50] Wang K, Chen G, Huang Z, Wan X, Huang FBridging the domain gap: Improve informal language translation via counterfactual domain adaptation. Proceedings of the AAAI Conference on Artificial Intelligence (AAAI’ 21), 2021Virtual Event, February 2–9, 2021

[51] Xu X, He H, Zhang H, Xu Y, He SUnsupervised domain adaptation via importance sampling. IEEE Transactions on Circuits and Systems for Video Technology, 2019, 30(12): 4688-4699.

[52] Zellinger W, Moser B A, Grubinger T, Lughofer E, Natschläger T, Saminger-Platz SRobust unsupervised domain adaptation for neural networks via moment alignment. Information Sciences, 2019, 483174-191.

[53] Zhang W, Zhao J, Qin G, Wang SCross-project defect prediction based on autoencoder with dynamic adversarial adaptation. Applied Intelligence, 2025, 55(5):324.

[54] Zhang N, Zhu K, Zhu DIAPCP: An effective cross-project defect prediction model via intra-domain alignment and programming-based distribution adaptation. IET Software, 2024, 2024(1):5358773.

[55] Zimmermann T, Nagappan N, Gall H, Giger E, Murphy BCross-project defect prediction: A large scale experiment on data vs. domain vs. process. Proceedings of the 7th Joint Meeting of the European Software Engineering, 2009August 24–28, 2009

[56] Zou Q, Lu L, Yang Z, Gu X, Qiu SJoint feature representation learning and progressive distribution matching for cross-project defect prediction. Information and Software Technology, 2021, 137106588.

Rights & permissions

Systems Engineering Society of China and Springer-Verlag GmbH Germany