BeeCF: Artificial bee colony based counterfactual explanations for healthcare recommendation
APPLIED SOFT COMPUTING, cilt.204, 2027 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 204
- Basım Tarihi: 2027
- Doi Numarası: 10.1016/j.asoc.2026.116444
- Dergi Adı: APPLIED SOFT COMPUTING
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, Compendex, INSPEC
- Gazi Üniversitesi Adresli: Evet
Özet
Counterfactual explanations have emerged as an effective approach for enhancing the interpretability of machine learning models, particularly in high-stakes domains such as healthcare. However, existing counterfactual gen eration methods often face challenges in producing valid and proximal explanations in a reliable and consistent manner. In this study, we propose a novel counterfactual explanation approach, named BeeCF, which employs the Artificial Bee Colony (ABC) algorithm as an optimization strategy. Specifically, the ABC algorithm is formu lated to solve an optimization problem that jointly maximizes the validity of the generated counterfactual while minimizing its proximity to the original instance. The BeeCF approach is evaluated on multiple publicly avail able healthcare datasets and compared against commonly used methods, including Random Sampling, Genetic Algorithm, and KD-tree-based Search, for counterfactual generation. In addition, we compare BeeCF against Differential Evolution (DE), a well known metaheuristic optimization method, to further validate its effective ness. Experimental results demonstrate that BeeCF consistently achieves high validity rates, approaching 100%, while generating counterfactuals with lower proximity and, in most cases, the highest plausibility scores among the compared methods. These findings highlight the effectiveness of ABC based optimization for counterfactual generation and its potential contribution to trustworthy healthcare recommendation systems.