An empirical study of bagging and boosting ensembles for identifying faulty classes in object-oriented software.
Hamoud AljamaanMahmoud O. ElishPublished in: CIDM (2009)
Keyphrases
- ensemble methods
- ensemble learning
- object oriented software
- decision stumps
- base classifiers
- base learners
- decision tree ensembles
- ensemble classifier
- rotation forest
- data flow
- ensemble classification
- decision trees
- random forests
- class labels
- design patterns
- object oriented
- negative correlation learning
- weighted voting
- prediction accuracy
- generalization ability
- machine learning methods
- multi class
- ensemble selection
- randomized trees
- majority voting
- benchmark datasets
- classifier ensemble
- gradient boosting
- ensemble members
- weak classifiers
- weak learners
- random forest
- training set
- imbalanced data
- software architecture
- neural network ensembles
- meta learning
- software development
- regression testing
- naive bayes
- high level
- multiple classifier systems
- boosting algorithms
- class distribution
- tree ensembles