A comparative study of iterative and non-iterative feature selection techniques for software defect prediction

Khoshgoftaar, Taghi M.; Gao, Kehan; Napolitano, Amri; Wald, Randall

doi:10.1007/s10796-013-9430-0

Cited by 69 publications

(40 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…By statistical analysis, they proved that under-sampling was useful in improving the prediction performance of the classifiers trained afterward. Khoshgoftaar et al [36] discussed the effects of random sampling combined with other data preprocessing methods (including feature ranking). Their results also proved the effectiveness of random sampling in dealing with imbalanced datasets.…”

Section: Related Workmentioning

confidence: 99%

“…But the purpose of the instance sampling was to reduce the total number of instances instead of handling class imbalance. Khoshgoftaar et al [36] combined filter-based feature ranking methods and random under-sampling. Their purpose was to use instance sampling for selecting features iteratively, but they did not train the classifier with balanced data.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Empirical Studies of a Two-Stage Data Preprocessing Approach for Software Fault Prediction

Liu

et al. 2016

IEEE Trans. Rel.

View full text Add to dashboard Cite

Software fault prediction is a valuable exercise in software quality assurance to best allocate limited testing resources. Classification is one of the effective methods for software fault prediction. The classification models are trained based on the datasets obtained by mining software historical repositories. However, the performance of the models depends on the quality of datasets. In this paper, we propose a novel two-stage data preprocessing approach which incorporates both feature selection and instance reduction. Specifically, in the feature selection stage, we first perform relevance analysis, and then propose a threshold-based clustering method, called novel threshold-based clustering algorithm, to conduct redundancy control. In the instance reduction stage, we apply random under-sampling to keep the balance between the faulty and non-faulty instances. In empirical studies, we chose datasets from real-world software projects, such as Eclipse and NASA. Then we compared our approach with some classical baseline methods, and further investigated the influencing factors in our approach. The final results demonstrate the effectiveness of our approach, and provide a guideline for achieving cost-effective data preprocessing when using our two-stage approach.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Empirical Studies of a Two-Stage Data Preprocessing Approach for Software Fault Prediction

Liu

et al. 2016

IEEE Trans. Rel.

View full text Add to dashboard Cite

show abstract

“…Their experimental results demonstrate that with the deep exploration of concept associations and information integration, the detection of high-level concepts from video shots can be further improved by a non-trivial margin. Khoshgoftaar et al (2014) target the problem of feature selection in a high dimensional feature space and conduct a comparative study of various feature selection techniques for software defect prediction in their paper "A comparative study of iterative and non-iterative feature selection techniques for software defect prediction." In particular, they try to tackle the problems of feature selection and imbalanced training data concurrently by using an iterative feature selection approach that iteratively does data sampling (to address the class imbalance issue) and feature selection (to address the curse of dimensionality issue).…”

Section: About This Special Issuementioning

confidence: 99%

Guest editorial: Information reuse, integration, and reusable systems

et al. 2014

View full text Add to dashboard Cite

“…It predicts whether a new module has a faultprone based on the software metrics and historical defect information. These defect information can not only be used to guide the allocation of resources and repair software defects timely, but also can save the cost of software development and improve the quality of software [1][2][3]. There are many kinds of software metrics used for defect prediction.…”

Section: Introductionmentioning

confidence: 99%

Software Defect Prediction Based on Data Sampling and Multivariate Filter Feature Selection

Lin¹,

Zhong²

2018

Proceedings of the 2018 2nd International Conference on Artificial Intelligence: Technologies and Applications (ICAITA 2018)

View full text Add to dashboard Cite

Abstract-In order to solve the useless feature and class imbalance problem in software defect prediction(SDP), this paper proposes a new prediction method which is based on data sampling and multivariate filter feature selection. Firstly, the sampling method re-samples the data set to achieve the data balance. Secondly, the multivariate filter algorithm selects feature and eliminates useless features such as irrelevant features and redundant features. Experimental results show that the proposed algorithm not only can effectively improve the prediction accuracy of the minority classes, but also effectively improve the overall classification performance of SDP.

show abstract

A comparative study of iterative and non-iterative feature selection techniques for software defect prediction

Cited by 69 publications

References 17 publications

Empirical Studies of a Two-Stage Data Preprocessing Approach for Software Fault Prediction

Empirical Studies of a Two-Stage Data Preprocessing Approach for Software Fault Prediction

Guest editorial: Information reuse, integration, and reusable systems

Software Defect Prediction Based on Data Sampling and Multivariate Filter Feature Selection

Contact Info

Product

Resources

About