Abstract | ||
---|---|---|
•Four ensemble learning strategies are proposed to predict change prone classes by aggregating fitness-based PSO classifiers.•The constituents of ensembles are aggregated using weighted voting (based on accuracy and ability to detect hard instances).•The empirical study uses ten popular open source data sets developed using Java language.•The proposed strategies of ensemble learning were found effective in predicting change-prone classes.•They were found better than individual constituent classifiers and competent to machine-learning ensembles. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1016/j.infsof.2018.05.007 | Information and Software Technology |
Keywords | Field | DocType |
Empirical validation,Ensemble learners,Particle swarm optimization,Software change prediction | Particle swarm optimization,Data mining,Voting,Computer science,Weighted voting,Fitness function,Software,Classifier (linguistics),Change prediction,Ensemble learning | Journal |
Volume | ISSN | Citations |
102 | 0950-5849 | 1 |
PageRank | References | Authors |
0.35 | 47 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ruchika Malhotra | 1 | 533 | 35.12 |
Megha Khanna | 2 | 59 | 6.47 |