A feature selection approach based on a similarity measure for software defect prediction

Yu, Qiao; Jiang, Shujuan; Wang, Rong-cun; Wang, Hongyang

doi:10.1631/fitee.1601322

Cited by 30 publications

(22 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Those features which are more relevant to the class and irrelevant to other features are preferred. Moreover, we apply three feature ranking approaches, including similarity measure (SM) [30], correlation (CL) [31] and gain ratio (GR) [32]. SM is designed to update the feature weights based on the similarity measure, and we can get a feature ranking list by sorting the feature weights in descending order.…”

Section: Cross-project Defect Prediction With Feature Selectionmentioning

confidence: 99%

An Empirical Study on the Effectiveness of Feature Selection for Cross-Project Defect Prediction

Qian

Jiang

et al. 2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

Software defect prediction has attracted much attention of researchers in software engineering. At present, feature selection approaches have been introduced into software defect prediction, which can improve the performance of traditional defect prediction (known as within-project defect prediction, WPDP) effectively. However, the studies on feature selection are not sufficient for cross-project defect prediction (CPDP). In this paper, we use the feature subset selection and feature ranking approaches to explore the effectiveness of feature selection for CPDP. An empirical study is conducted on NASA and PROMISE datasets. The results show that both the feature subset selection and feature ranking approaches can improve the performance of CPDP. Therefore, we should select the representative feature subset or set a reasonable proportion of selected features to improve the performance of CPDP in future studies. INDEX TERMS Software defect prediction, cross-project defect prediction, feature selection, feature ranking.

show abstract

Section: Cross-project Defect Prediction With Feature Selectionmentioning

confidence: 99%

An Empirical Study on the Effectiveness of Feature Selection for Cross-Project Defect Prediction

Qian

Jiang

et al. 2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…Rodriguez et al (2014) proposed a key for defect removal and defect prediction. He made a list of metrics first and then used it for software defect prediction [4,6]. Elish et al (2008) used the Naï ve Bayes classifier for predicting software defects and achieved 71 percent accuracy on fault detection [7][8].…”

Section: Related Work Donementioning

confidence: 99%

“…Al (2010) and Gray et al( 2011) proposed several defect prediction models using classical available techniques in machine learning with MDP datasets. However, they ignored the class imbalance issue, making their results biased and erroneous [4][5]. Adopting a systematic approach, we propose a novel defect prediction system using feature selection for removing noisy and irrelevant attributes and entropy for the class imbalance problem.…”

Section: Introductionmentioning

confidence: 99%

Proposed Intelligent Software System for Early Fault Detection

Banga¹,

Bansal²,

Singh³

2019

Int J Performability Eng

View full text Add to dashboard Cite

The major challenge in designing an Intelligent Information Software System for fault detection is to detect faults at an early stage unless it becomes a failure. This can be achieved by using feature selection and effective classification applied on failure datasets. Support Vector Machines (SVM) are used for efficient and accurate feature selection by finding unknown model parameters using local and global kernel parameter optimization. Research shows that previously many attempts were made using classical classifiers as decision tree, naï ve bayes, and k-NN for software fault prediction. In earlier research, class imbalance problems in software fault datasets were not addressed. In this paper, we propose an intelligent hybrid algorithm that is based on feature selection hybrid kernel function SVM and entropy-based bagging for efficient classification to reduce the class imbalance problem. The proposed model is compared with traditional approaches. The improved hybrid algorithm based on entropy-based bagging and mixed kernel SVM can effectively improve the classification accuracy of NASA Metric Data Program (MDP) faulty datasets. This paper presents an empirical study on using the proposed hybrid algorithm and results showed that our proposed approach enhances the classification accuracy when compared with existing methods.

show abstract

“…The authors' approach overlooked this in their approach. In [12], Qiao et al proposed a feature weighted approach based on a distance measure for software defect prediction. The authors designed a ranking algorithm that maintains a rank list of the features.…”

Section: Related Workmentioning

confidence: 99%

“…The work in [11] is particularly noteworthy because of the authors' contribution to developing a baseline experiment that most researchers followed since. This allowed a direct comparison of the different approaches proposed in literature [12], [13]. Six issues with existing state-of-the-art techniques were discussed in [14] and while there have been a few papers that try to address the issues raised in the paper, some of the core problems still remain.…”

Section: Introductionmentioning

confidence: 99%

Modeling Software Defects as Anomalies: A Case Study on Promise Repository

Neela¹,

Asif²,

Ami³

et al. 2017

JSW

View full text Add to dashboard Cite

Software defect prediction is a highly studied domain in Software Engineering research due to its importance in software development. In literature, various classification methods with static code attributes have been used to predict defects. However, defected instances are very few compared to non-defected instances and as such lead to imbalanced data. Traditional machine learning techniques give poor results for such data. In this paper an anomaly detection technique for software defect prediction, is proposed which is not affected by imbalanced data. The technique incorporates both univariate and multivariate Gaussian distribution to model non-defected software module. The defected software modules are then predicted based on their deviation from the generated model. To evaluate our approach, we implemented the algorithm and tested it on the NASA datasets from the PROMISE repository. By utilizing this approach, we observed an average balance of 63.36% and 69.06% in univariate model and multivariate model respectively. Without utilizing optimization or filter, this approach yield better result than industry standard of 60%.

show abstract

A feature selection approach based on a similarity measure for software defect prediction

Cited by 30 publications

References 19 publications

An Empirical Study on the Effectiveness of Feature Selection for Cross-Project Defect Prediction

An Empirical Study on the Effectiveness of Feature Selection for Cross-Project Defect Prediction

Proposed Intelligent Software System for Early Fault Detection

Modeling Software Defects as Anomalies: A Case Study on Promise Repository

Contact Info

Product

Resources

About