Software Defect Prediction Using Process Metrics Systematic Literature Review: Dataset and Granularity Level

Ramadhina, Salsabila; Bahawares, Rizal Broer; Hermadi, Irman; Suroso, Arif Imam; Rodoni, Ahmad; Arkeman, Yandra

doi:10.1109/citsm52892.2021.9587932

Cited by 5 publications

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Over time many different metrics have been proposed. Process metrics [41], [42], [43], [44], [45], change metrics [46], [47], [48], [49], semantic metrics [24], [25], [26], [27] and others.…”

Section: B Metrics Prediction Granularity and Approaches To Sdpmentioning

confidence: 99%

Empirical Study: How Issue Classification Influences Software Defect Prediction

et al. 2023

View full text Add to dashboard Cite

Software defect prediction aims to identify potentially defective software modules to better allocate limited quality assurance resources. Practitioners often do this by utilizing supervised models trained using historical data. This data is gathered by mining version control and issue tracking systems. Version control commits are linked to issues they address. If the linked issue is classified as a bug report, the change is considered as bug fixing. The problem arises from the fact that issues are often incorrectly classified within issue tracking systems. This introduces noise into the gathered datasets. In this paper, we investigate the influence issue classification has on software defect prediction dataset quality and resulting model performance. To do this, we mine data from 7 popular open-source repositories, create issue classification and software defect prediction datasets for each of them. We investigate issue classification using four different methods; a simple keyword heuristic, an improved keyword heuristic, the FastText model and the RoBERTa model. Our results show that using the RoBERTa model for issue classification produces the best software defect prediction datasets, containing on average 14.3641% of mislabeled instances. SDP models trained on such datasets achieve superior performance, to those trained on SDP datasets created using other issue classification methods, in 65 out of 84 experiments, with 55 of them being statistically relevant. Furthermore, in 17 out of 28 experiments we could not show a statistically relevant performance difference between SDP models trained on RoBERTa derived software defect prediction datasets and those created using manually labeled issues.

show abstract

Section: B Metrics Prediction Granularity and Approaches To Sdpmentioning

confidence: 99%

Empirical Study: How Issue Classification Influences Software Defect Prediction

et al. 2023

View full text Add to dashboard Cite

show abstract

“…In literature, several performance metrics, such as project metrics, 24,25 product metrics, 26,27 and process metrics, 28,29 are used to analyze a project from different perspectives (cost, effort, time, size, functionality, capability, reliability, etc.). However, none of them examined the system using bad smells and software quality metrics which are two important aspects of code that describes the system's quality.…”

Section: Motivationmentioning

confidence: 99%

Severity Factor (SF): An aid to developers for application of refactoring operations to improve software quality

Agnihotri

Chug

2023

J Software Evolu Process

View full text Add to dashboard Cite

Bad smells are certain flaws in the structure of the code that might not disturb the normal functioning of a program but negatively affects the software quality. Developers use refactoring as a corrective measure for the treatment of bad smells. The current study aids the developers in the application of refactoring by identifying the critical classes, that is, classes that are challenging to maintain and are of degraded quality. In this study, 10 quality metrics and 10 bad smells have been selected to conduct an investigation on different releases of five open‐source systems. A new metric, severity factor (SF) has been introduced that categorizes the classes of the selected systems into four criticality levels—severe, major, mid, and low. Also, the relationship between SF, criticality levels, and the refactoring operations has been analyzed. The findings show that 60% of the total classes have been affected by bad smells, and long statement is the most dominant smell present in 27.6% of the classes. The results show 84% of the refactoring operations have been performed on highly critical classes. Thus, the SF metric plays a crucial role in driving the developer's attention to the critical classes that need to be treated urgently.

show abstract