EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

An Optimized Hybrid Artificial Immune System for Breast Cancer Diagnosis Using Leakage-Free Cross-Validation

Maen MARWAN Alzubi

Journal of Robotics and Control (JRC) · 01.03.2026 · 7 (2) · 3658-3670

DOI: 10.18196/jrc.v7i2.29388 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Robotics and Control (JRC), online erschienen am 01.03.2026. Verfasst von Maen MARWAN Alzubi. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Breast cancer diagnosis from structured numerical features remains an important binary classification problem in medical machine learning, particularly when high sensitivity and low false-negative rates are required. This study proposes an Optimized Hybrid Artificial Immune System for benign–malignant breast cancer classification using the Wisconsin Diagnostic Breast Cancer dataset. The research contribution is an optimized hybrid diagnostic framework that integrates an enhanced Clonal Selection Algorithm with Support Vector Machine and Random Forest classifiers through a transparent probability-level fusion strategy. The optimized Clonal Selection Algorithm uses class-stratified antibody initialization, bagging-based population generation, k-nearest-neighbor antibody voting, elitism, margin-based fitness evaluation, and adaptive cosine mutation. The dataset contains 569 instances and 30 numerical features, with 357 benign and 212 malignant cases. To avoid data leakage, MinMax normalization was fitted only on the training partition within each fold, and model performance was evaluated using stratified 10-fold cross-validation. The proposed Optimized Hybrid AIS achieved an accuracy of 0.9789 ± 0.0138, precision of 0.9816 ± 0.0238, recall of 0.9621 ± 0.0375, F1-score of 0.9712 ± 0.0194, ROC-AUC of 0.9950 ± 0.0077, and MCC of 0.9554 ± 0.0296. The pooled confusion matrix showed 204 true positives, 353 true negatives, 4 false positives, and 8 false negatives. Statistical testing demonstrated significant improvements over Random Forest, k-nearest neighbors, Naive Bayes, Gradient Boosting, Decision Tree, Linear Discriminant Analysis, XGBoost, and the standalone Optimized CSA, while the differences against SVM and Neural Network were not statistically significant. These findings support the proposed model as a strong computational benchmark for breast cancer diagnosis, while external validation remains necessary.

Zitieren

APA

Alzubi, M. M. (2026). An Optimized Hybrid Artificial Immune System for Breast Cancer Diagnosis Using Leakage-Free Cross-Validation. Journal of Robotics and Control (JRC), 7(2), 3658-3670. https://doi.org/10.18196/jrc.v7i2.29388

Vancouver

Alzubi MM. An Optimized Hybrid Artificial Immune System for Breast Cancer Diagnosis Using Leakage-Free Cross-Validation. Journal of Robotics and Control (JRC). 2026;7(2):3658-3670. doi:10.18196/jrc.v7i2.29388

BibTeX
@article{Alzubi2026,
  author = {Alzubi, Maen MARWAN},
  title = {An Optimized Hybrid Artificial Immune System for Breast Cancer Diagnosis Using Leakage-Free Cross-Validation},
  journal = {Journal of Robotics and Control (JRC)},
  year = {2026},
  volume = {7},
  number = {2},
  pages = {3658--3670},
  doi = {10.18196/jrc.v7i2.29388}
}
Kostenlos starten →