Back to Search
Start Over
Exclusive use and evaluation of inheritance metrics viability in software fault prediction—an experimental study
- Source :
- PeerJ Computer Science, PeerJ Computer Science, Vol 7, p e563 (2021)
- Publication Year :
- 2021
- Publisher :
- PeerJ Inc., 2021.
-
Abstract
- Software Fault Prediction (SFP) assists in the identification of faulty classes, and software metrics provide us with a mechanism for this purpose. Besides others, metrics addressing inheritance in Object-Oriented (OO) are important as these measure depth, hierarchy, width, and overriding complexity of the software. In this paper, we evaluated the exclusive use, and viability of inheritance metrics in SFP through experiments. We perform a survey of inheritance metrics whose data sets are publicly available, and collected about 40 data sets having inheritance metrics. We cleaned, and filtered them, and captured nine inheritance metrics. After preprocessing, we divided selected data sets into all possible combinations of inheritance metrics, and then we merged similar metrics. We then formed 67 data sets containing only inheritance metrics that have nominal binary class labels. We performed a model building, and validation for Support Vector Machine(SVM). Results of Cross-Entropy, Accuracy, F-Measure, and AUC advocate viability of inheritance metrics in software fault prediction. Furthermore, ic, noc, and dit metrics are helpful in reduction of error entropy rate over the rest of the 67 feature sets.
- Subjects :
- General Computer Science
Computer science
02 engineering and technology
computer.software_genre
Software reliability
Software testing
Inheritance (object-oriented programming)
Software
Artificial Intelligence
Machine learning
0202 electrical engineering, electronic engineering, information engineering
Feature (machine learning)
Preprocessor
Hierarchy (mathematics)
Software inheritance metrics
business.industry
Software fault prediction
Software Engineering
020207 software engineering
QA75.5-76.95
Software metric
Software quality
Support vector machine
Algorithms and Analysis of Algorithms
Electronic computers. Computer science
020201 artificial intelligence & image processing
Data mining
business
Software metrics
computer
Subjects
Details
- Language :
- English
- ISSN :
- 23765992
- Volume :
- 7
- Database :
- OpenAIRE
- Journal :
- PeerJ Computer Science
- Accession number :
- edsair.doi.dedup.....40bb070365260c4f32091966c964574e