Learning from Imbalanced Data for Predicting the Number of Software Defects

Yu, Xiao; Liu, Jin; Yang, Zijiang; Jia, Xiangyang; Qi, Ling; Ye, Sizhe

doi:10.1109/issre.2017.18

Cited by 37 publications

(23 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then found that using decision tree regression can achieve the highest performance in both within-project prediction scenario and cross-project prediction scenario. Yu et al [12] explored resampling (i.e., SMOTEND and RUSND) and ensemble learning (i.e., AdaBoost.R2) methods. Then they proposed two hybrid methods (i.e., SMOTENDBoost and RUSNDBoost) and these two methods can achieve higher performance.…”

Section: Related Workmentioning

confidence: 99%

“…In our empirical studies, 30 projects from PROMISE are used to verify the performances of MPR. These projects can be downloaded from PROMISE and they are widely used in previous empirical studies [11,12,15,16,21,22]. The characteristic of these projects are shown in Table 1, which includes project name, number of modules, number (percentage) of defective modules and the maximum defects contained in the modules.…”

Section: Experimental Subjectsmentioning

confidence: 99%

See 1 more Smart Citation

Multi-project Regression based Approach for Software Defect Number Prediction

Huang

Chen

et al. 2019

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

Software defect prediction can make software quality assurance (SQA) process more efficient, economic and targeted. Previous studies mainly focused on classifying software modules as defect-prone or not. However, prediction the number of defects for a new software module is rarely investigated. Moreover, these studies built models independently for each project, which may ignore the relatedness among multiple projects. To effectively utilize the relatedness, we propose a novel approach MPR (multiproject regression) for SDNP (software defect number prediction). To verify the effectiveness of MPR, we perform experimental studies on 30 real-world projects and compare our approach with 6 state-of-the-art baselines (i.e., LR, NNR, SVR, DTR, BRR and DBR). AAE (Average absolute error) and ARE (average relative error) performance measures are used to evaluate the performance of MPR. The results show MPR can achieve better performance in most cases, which indicates the competitiveness of MPR in the context of SDNP.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Experimental Subjectsmentioning

confidence: 99%

Multi-project Regression based Approach for Software Defect Number Prediction

Huang

Chen

et al. 2019

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

show abstract

“…Fault prediction aims to predict whether or not a particular software module is defective via learning from historical defect data [1], [2]. Therefore, fault prediction is often used to help to reasonably allocate limited development and maintenance resources [3]- [5]. Many learning models have been proposed for fault prediction.…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Feature Selection Method for Software Fault Prediction

Jian

Zhou

et al. 2019

IEICE Trans. Inf. & Syst.

Self Cite

View full text Add to dashboard Cite

Fault prediction aims to identify whether a software module is defect-prone or not according to metrics that are mined from software projects. These metric values, also known as features, may involve irrelevance and redundancy, which hurt the performance of fault prediction models. In order to filter out irrelevant and redundant features, a Hybrid Feature Selection (abbreviated as HFS) method for software fault prediction is proposed. The proposed HFS method consists of two major stages. First, HFS groups features with hierarchical agglomerative clustering; second, HFS selects the most valuable features from each cluster to remove irrelevant and redundant ones based on two wrapper based strategies. The empirical evaluation was conducted on 11 widely-studied NASA projects, using three different classifiers with four performance metrics (precision, recall, F-measure, and AUC). Comparison with six filter-based feature selection methods demonstrates that HFS achieves higher average F-measure and AUC values. Compared with two classic wrapper feature selection methods, HFS can obtain a competitive prediction performance in terms of average AUC while significantly reducing the computation cost of the wrapper process.

show abstract

“…So far, many efficient software defect prediction methods using statistical methods or machine learning techniques have been proposed [2][3][4], but they are usually confined to predicting a given software module being faulty or non-faulty by means of some binary classification techniques. 1 However, predicting the defect-prone of a given software module does not provide enough logistics to software testing in practice [5][6]. Some of the faulty software modules may have comparatively vast quantities of faults compared to other 1 DOI reference number: 10.18293/SEKE2018-068 modules and hence require some additional maintenance resources to fix them.…”

Section: Introductionmentioning

confidence: 99%

“…If we are able to predict the accurate number of faults, software testers will pay particular attention to those software modules that have more number of faults, which makes testing processes more efficient in the case of limited development and maintenance resources. Thus, predicting the number of faults in software modules can be more helpful instead of predicting the modules being faulty or non-faulty [6].…”

Section: Introductionmentioning

confidence: 99%

Revisting the Impact of Regression Models for Predicting the Number of Defects

Wu¹,

Ye²,

Li³

et al. 2018

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

Predicting the number of faults in software modules can be more helpful instead of predicting the modules being faulty or non-faulty. Chen et al. (SEKE 397-402, 2015) and Rathore et al. (Soft Computing 21: 7417-7434, 2017) empirically investigate the feasibility of some regression algorithms for predicting the number of defects. The experimental results showed that the decision tree regression algorithm performed best in terms of average absolute error (AAE), average relative error (ARE) and root mean square error (RMSE). However, they did not consider the imbalanced data distribution problem in defect datasets and employed improper performance measures for evaluating the regression models to evaluate the performance of models for predicting the number of defects. Hence, we revisit the impact of different regression algorithms for predicting the number of defects using Fault-Percentile-Average (FPA) as the performance measure. The experiments on 31 datasets from PROMISE repository show that the prediction performance of models for predicting the number of defects built by different regression algorithms are various, and the gradient boosting regression algorithm and the Bayesian ridge regression algorithm can achieve better performance．

show abstract

Learning from Imbalanced Data for Predicting the Number of Software Defects

Cited by 37 publications

References 63 publications

Multi-project Regression based Approach for Software Defect Number Prediction

Multi-project Regression based Approach for Software Defect Number Prediction

A Hybrid Feature Selection Method for Software Fault Prediction

Revisting the Impact of Regression Models for Predicting the Number of Defects

Contact Info

Product

Resources

About