TLEL: A two-layer ensemble learning approach for just-in-time defect prediction

Yang, Xinli; Lo, David; Xia, Xin; Sun, Jifei

doi:10.1016/j.infsof.2017.03.007

Cited by 198 publications

(132 citation statements)

References 23 publications

Supporting

Mentioning

121

Contrasting

Unclassified

Order By: Relevance

“…Kamei's features contain 14 features in total, which have been proposed and validated by Kamei et al and used by Yang et al, and Yang et al in change classification. Kamei's features are divided into diffusion, size, purpose, history, and experience dimensions.…”

Section: Change Classification Methodologymentioning

confidence: 99%

“…Two overall measures F1‐measure and area under curve (AUC) are also used to consider the overall performance for both buggy changes and clean changes. What is more, cost effectiveness has become widely used in bug prediction and in which PofB20 is the most popular one . For the aforementioned reasons, we adopt five performance measures totally, including buggy recall, buggy precision, buggy F1‐measure, AUC, and PofB20.…”

Section: Empirical Studymentioning

confidence: 99%

“…When J48 is used as the base classifier, only one tree is built, whereas when RF is used, 10 trees are built according to the default values in Weka. Since ensemble learning uses the combination of several base classifiers to make the final decision, we wonder if further combining several trees in the base classifier, as is done by Yang et al, will provide better performance. The comparison results are shown in Table .…”

Section: Empirical Studymentioning

confidence: 99%

“…Tan et al, Kamei et al, and Yang et al have paid attention to the imbalance problem of change classification. Tan et al adopted four resampling approaches, including simple duplicate, synthetic minority oversampling technique, spread subsample, and resampling with/ without replacements, to alleviate the class imbalance problem.…”

Section: Introductionmentioning

confidence: 99%

“…It is still not clear which kind(s) of these features are better. Furthermore, the prediction methods of Yang et al have shown to be better than the others if extension of their change features with the other three kinds used by Kim et al can yield better results. The exploration to this question can give the answer.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

An empirical study of software change classification with imbalance data‐handling methods

et al. 2018

View full text Add to dashboard Cite

Summary Bug prediction in software code changes can help developers to find out and fix bugs immediately when they are introduced, thus to improve the effectiveness and validity of bug fixing. In data mining, this problem can be regarded as a change classification task. However, one of its key characteristics, ie, class‐imbalance, holds back the performance of standard classification methods. In this paper, we consider a quantity of imbalance data‐handling methods and extract a more comprehensive groups of change features, aiming to achieve better change classification performance. Two different types of imbalance data‐handling methods, namely, resampling and ensemble learning methods, are employed. Especially, we explore the performance of their combination. To compare the performance of different imbalance data‐handling methods, an experiment with 10 open source projects is conducted. Four classification methods, including J48, Naïve Bayes, SMO, and Random Forest, are used as standard classifiers and as the base classifiers, respectively. Moreover, contribution of different groups of change features are evaluated. Experimental results show that imbalance data‐handling methods can improve the performance of change classification and the combination methods, which take advantage of both ensemble learning and resampling, perform better than using ensemble learning methods or resampling methods individually. Of the studied imbalance data‐handling methods, the combination of Bagging and random undersampling with J48 as the base classifier yields out better prediction results than those achieved by other methods. Additionally, of the collected change features, text vector features accounts for a larger proportion than others.

show abstract

Section: Change Classification Methodologymentioning

confidence: 99%

Section: Empirical Studymentioning

confidence: 99%

Section: Empirical Studymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

An empirical study of software change classification with imbalance data‐handling methods

et al. 2018

View full text Add to dashboard Cite

show abstract

IM_NRFixer: A hybrid approach to alleviate class‐imbalance problem for predicting the fixability of Non‐Reproducible bugs

Goyal

Sardana

2020

J Software Evolu Process

View full text Add to dashboard Cite

Software maintenance is an important phase in the software development life cycle. Software projects maintain bug repositories to gather, organize, and keep track of bug reports. These bug reports are resolved by numerous software developers. Whenever the reported bug does not get resolved by the assigned developer, he marks the resolution of bug report as Non‐Reproducible (NR). When NR bugs are reconsidered, few of them get resolved, and their resolution changes from NR to fix (NRF). The main aim of this paper is to predict these fixable NRF bug reports. A major challenge in predicting NRF bugs from NR bugs is that only a small portion of NR bugs get fixed, i.e., class‐imbalance problem. For example, NRF bugs account for only 8.64%, 4.73 %, 4.56%, and 1.06% in NetBeans, Eclipse, Open Office, and Mozilla Firefox projects respectively. In this paper, we work on improving the classification performance on these imbalanced datasets. We propose IMNRFixer, a novel and hybrid NRF prediction tool. IMNRFixer uses three different techniques to combat class‐imbalance problem: undersampling, oversampling, and ensemble models. We evaluate the performance of IMNRFixer models on four large and open‐source projects of Bugzilla repository. Our results show that IMNRFixer outperforms conventional machine learning techniques. IMNRFixer achieves performance up to 71.7%, 93.1%, 91.7%, and 96.5% while predicting the minority class (NRF) for NetBeans, Eclipse, Open Office, and Mozilla Firefox projects, respectively.

show abstract

Just‐in‐time defect prediction for software hunks

et al. 2021

View full text Add to dashboard Cite

Just‐in‐time defect prediction can remind software developers and managers to verify and fix bugs at the moment they appeared, thus improving the effectiveness and validity of bug fixing. Existing studies mainly focus on just‐in‐time prediction for software files (JIT‐F). JIT‐F is a binary classification problem, which classifies (hence predicts) a file change as buggy or clean. This article provides a detailed analysis of just‐in‐time defect prediction for software hunks (JIT‐H), which predicts bugs at a finer level of granularity, and hence further improves the efficiency of bug fixing. Classification is performed using the ensemble technique of bagging—aggregated combinations of random under sampling plus multiple classifiers (J48 and Random Forest). An empirical study with 10 open source projects was conducted to validate the effectiveness of JIT‐H. Experimental results show that JIT‐H is effective at predicting defects in software hunk changes. Compared with JIT‐F, JIT‐H is more cost effective. Additionally, analysis on the change features indicates that Text Vector features and hunk change level features are of more importance than features in other groups and levels.

show abstract

TLEL: A two-layer ensemble learning approach for just-in-time defect prediction

Cited by 198 publications

References 23 publications

An empirical study of software change classification with imbalance data‐handling methods

An empirical study of software change classification with imbalance data‐handling methods

IM_NRFixer: A hybrid approach to alleviate class‐imbalance problem for predicting the fixability of Non‐Reproducible bugs

Just‐in‐time defect prediction for software hunks

Contact Info

Product

Resources

About

TLEL: A two-layer ensemble learning approach for just-in-time defect prediction

Cited by 198 publications

References 23 publications

An empirical study of software change classification with imbalance data‐handling methods

An empirical study of software change classification with imbalance data‐handling methods

IMNRFixer: A hybrid approach to alleviate class‐imbalance problem for predicting the fixability of Non‐Reproducible bugs

Just‐in‐time defect prediction for software hunks

Contact Info

Product

Resources

About

IM_NRFixer: A hybrid approach to alleviate class‐imbalance problem for predicting the fixability of Non‐Reproducible bugs