.png){kind=icon}
.png){kind=icon}
Return
Bagging-based ensemble classifiers using multi-objective genetic programming
.png){kind=icon}
PRE
AI
DOI：10.1007/s12293-025-00444-8.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
As an effective evolutionary computation algorithm, genetic programming (GP) can be designed as effective classifiers due to its flexible representation method. However, the classification performance of GP classifiers can be degraded due to imbalanced data and weak generalization ability. Precision-recall curve (PRC) has been proven to be an effective evaluation metric for dealing with imbalanced data. However, PRC may result in classifiers with the same PRC value being completely different classifiers. Moreover, controlling the complexity of GP individuals can improve their generalization ability. Therefore, in this paper, multi-objective GP (MOGP) is used to optimize three objectives including recall, precision and model complexity to reduce the impact of imbalanced data and improve the generalization of GP individuals. MOGP-based ensemble classifier construction methods can improve the generalization ability of classification models. However, this strategy needs to address the issues of how to improve the diversity of GP solutions and select optimal solutions from Pareto fronts. Therefore, in this paper, a bagging-based ensemble classifier construction method is proposed to improve the generalization of GP classifiers, which uses non-repeated sampling to generate multiple training subsets and runs MOGP multiple times on these training subsets to construct ensembles. Experiments on ten datasets show that our MOGP-based classifier construction method can achieve better classification performance than single-objective GP classifier construction methods, and our bagging-based ensemble classifier construction methods can further improve the classification performance compared to only using MOGP. Comparisons with six state-of-the-art GP classifier construction methods and six traditional machine learning algorithms show that our proposed approach can achieve significantly better classification performance in most cases.
Keywords:
Genetic programming
Multi-objective
Ensemble
Classification
Bagging
Journal
IF:
2.3
Papers:
452
Citations:
718