.png){kind=icon}
Return
Meta-heuristic optimization algorithm for predicting software defects
.png){kind=icon}
PRE
AI
DOI：10.1111/exsy.12768.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Software engineering companies strive to improve software quality by predicting software defects-prone modules. Although various data mining methods have been developed, unstable accuracy rates are still critical issues owing to the imbalanced nature and high dimensionality of software defect datasets. To deal with this issue, we propose a spotted hyena, a novel meta-heuristic optimization algorithm for predicting software defects. Support and confidence in classification rules are the basis of a multi-objective fitness function that assists the spotted hyena algorithm in serving as a classifier by finding the fittest classification or standard rules among individuals. Experiments were conducted on four NASA software datasets, JM1, KC2, KC1, and PC3. The spotted hyena classifier provides an accuracy of 85.2, 84, 89.6, and 81.8%, respectively, for these datasets. These accuracy rates are better than those achieved using other popular data mining techniques. We also discuss other classification measures in connection with the experimental results, such as precision, recall, and confusion matrices, in connection with the experimental results. Moreover, the Gaussian mixture model is used to study the uncertainty quantification of the proposed classifier. The study proved the feasible performance of the spotted hyena classifier in four different case studies.
Keywords:
confidence
software defect prediction
software metric
spotted hyena optimizer algorithm
support
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2.3
Papers:
2.6K
Citations:
3.8K
Organization
Cited Papers
Gaussian Mixture Based Uncertainty Modeling to Optimize Energy Management of Heterogeneous Building Neighborhoods: A Case Study of a Dutch University Medical Campus.png)
ENERGY AND BUILDINGS
OAAI
ENERGY AND BUILDINGSIF7.1
OAAIShareSaveShareSave
A review of uncertainty quantification in deep learning: Techniques, applications and challengesINFORMATION FUSIONOAAI
INFORMATION FUSIONIF15.5
OAAIShareSaveShareSaveShareSaveShareSave
View more