|
|
Bimonthly Since 1986 |
ISSN 1004-9037
|
|
|
|
|
Publication Details |
Edited by: Editorial Board of Journal of Data Acquisition and Processing
P.O. Box 2704, Beijing 100190, P.R. China
Sponsored by: Institute of Computing Technology, CAS & China Computer Federation
Undertaken by: Institute of Computing Technology, CAS
Published by: SCIENCE PRESS, BEIJING, CHINA
Distributed by:
China: All Local Post Offices
|
|
|
|
|
|
|
|
|
|
Abstract
Boosting approaches have recently been developed in the area of software defect prediction (SDP) by combining various base classifiers. The use of boosting has sometimes proven to be more accurate than using single base classifier in some experiments. Massive research from the industry and experts has started working on this area, even though most of the predictive methods are still in their infancy and require further research. To determine whether boosting models are superior to employing single classification models to produce high-quality software, more research is necessary. A Weighted Adaptive Boosting Ensemble (WABE) approach has been proposed for software defect prediction with software quality. The misclassification costs are incorporated into the weight-update rule of the boosting method, which causes the proposed algorithms to increase the weights of the samples linked to misclassified defect-prone modules. The proposed model showed that a high area under the receiver operating curve (ROC-AUC) and the learning ratio of the new model look promising, and various performance metrics are compared with other state-of-the-art machine learning-based methods to prove its superiority. This research confirms that decision tree classifiers must be carefully chosen as estimators to accurately identify the defective parts for an effective quality product.
Keyword
Ensemble Learning, Software defect, Software quality, Software Engineering, Weighted Adaptive Boosting.
PDF Download (click here)
|
|
|
|
|