Data Transformation in Cross-project Defect Prediction

Zhang, Feng; Keivanloo, Iman; Zou, Ying

doi:10.1007/s10664-017-9516-2

Cited by 56 publications

(48 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We use the McNemar test with the 95% confidence level to perform diversity analysis on defective modules between different CPDP methods. This diversity analysis method has also been used by Zhang et al in their previous study. The McNemar test is a nonparametric test, and it does not need the assumptions on the distribution of a subject variable.…”

Section: Methodsmentioning

confidence: 99%

“…The objective of metric value transformation is to make the transformed values of the metric satisfy the normal distribution assumption. Commonly used transformation methods include logarithmic transformation, min‐max normalization, z score normalization, rank transformation, or Box‐Cox transformation . (b) Some researchers focus on selecting source projects, which are more similar to the target project.…”

Section: Background and Related Workmentioning

confidence: 99%

“…However, they found there exists difference between these classifiers in terms of the specific defective modules each detects and does not detect. Some researchers performed diversity analysis for cross‐project defect prediction . However, they only performed diversity analysis in view of metric transformation methods or classifiers.…”

Section: Background and Related Workmentioning

confidence: 99%

“…For example, since different software metrics rarely follow a normal distribution, Zhang et al investigated whether different transformation methods (ie, log transformation, rank transformation, and Box‐Cox transformation) cause distinct predictions on the same defective module. While Panichella et al compared six different classifiers (ie, logistic regression, RBFNetwork, ADTree, decision table, multilayer perception, and Bayes Network) in CPDP and draw the same conclusion in Zhang et al . However, to the best of our knowledge, no researchers have investigated whether different state‐of‐the‐art CPDP methods can find the same defective modules.…”

Section: Background and Related Workmentioning

confidence: 99%

“…However, to the best of our knowledge, no researchers have investigated whether different state‐of‐the‐art CPDP methods can find the same defective modules. Therefore, we consider eight supervised CPDP methods and four unsupervised CPDP methods, which have not been investigated in previous studies and then perform diversity analysis on five widely used SDP datasets.…”

Section: Background and Related Workmentioning

confidence: 99%

See 4 more Smart Citations

Do different cross‐project defect prediction methods identify the same defective modules?

Chen

Qu³

et al. 2019

J Software Evolu Process

View full text Add to dashboard Cite

Cross‐project defect prediction (CPDP) is needed when the target projects are new projects or the projects have less training data, since these projects do not have sufficient historical data to build high‐quality prediction models. The researchers have proposed many CPDP methods, and previous studies have conducted extensive comparisons on the performance of different CPDP methods. However, to the best of our knowledge, it remains unclear whether different CPDP methods can identify the same defective modules, and this issue has not been thoroughly explored. In this article, we select 12 state‐of‐the‐art CPDP methods, including eight supervised methods and four unsupervised methods. We first compare the performance of these methods in the same experiment settings on five widely used datasets (ie, NASA, SOFTLAB, PROMISE, AEEEM, and ReLink) and rank these methods via the Scott‐Knott test. Final results confirm the competitiveness of unsupervised methods. Then we perform diversity analysis on defective modules for these methods by using the McNemar test. Empirical results verify that different CPDP methods may lead to difference in the modules predicted as defective, especially when the comparison is performed between the supervised methods and unsupervised methods. Finally, we also find there exist a certain number of defective modules, which cannot be correctly identified by any of the CPDP methods or can be correctly identified by only one CPDP method. These findings can be utilized to design more effective methods to further improve the performance of CPDP.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

See 3 more Smart Citations