English

Building Defect Prediction Models by Online Learning Considering Defect Overlooking

Software Engineering 2024-04-18 v1

Abstract

Building defect prediction models based on online learning can enhance prediction accuracy. It continuously rebuilds a new prediction model, when a new data point is added. However, a module predicted as "non-defective" can result in fewer test cases for such modules. Thus, a defective module can be overlooked during testing. The erroneous test results are used as learning data by online learning, which could negatively affect prediction accuracy. To suppress the negative influence, we propose to apply a method that fixes the prediction as positive during the initial stage of online learning. Additionally, we improved the method to consider the probability of the overlooking. In our experiment, we demonstrate this negative influence on prediction accuracy, and the effectiveness of our approach. The results show that our approach did not negatively affect AUC but significantly improved recall.

Keywords

Cite

@article{arxiv.2404.11033,
  title  = {Building Defect Prediction Models by Online Learning Considering Defect Overlooking},
  author = {Nikolay Fedorov and Yuta Yamasaki and Masateru Tsunoda and Akito Monden and Amjed Tahir and Kwabena Ebo Bennin and Koji Toda and Keitaro Nakasai},
  journal= {arXiv preprint arXiv:2404.11033},
  year   = {2024}
}

Comments

4 pages, 4 figures, 4 tables. The paper is currently under review in IEICE. arXiv admin note: text overlap with arXiv:2308.13582

R2 v1 2026-06-28T15:56:40.164Z