Robust Breast Cancer Classification Using Ensemble Feature Selection And Weighted Stability Regularization

Manmohan sahoo; Aswini kumar mohanty

doi:10.29284/tb9nxq10

Authors

Manmohan sahoo Research scholar, Biju Patnaik University of Technology Rourkela Odisha.
Aswini kumar mohanty Principal, synergy Institute of technology Phulnakhara, Bhubaneswar,

DOI:

https://doi.org/10.29284/tb9nxq10

Keywords:

Breast cancer diagnosis, hybrid feature selection, predictive modeling, SHAP values, Random Forest classifier, XGBoost algorithm.

Abstract

Detecting breast cancer early can significantly improve patient survival and treatment outcomes. However, the large number of features in medical imaging data makes it challenging to identify the most relevant predictors. This study introduces a framework that strengthens breast-cancer diagnosis by combining several complementary feature-selection methods—Boruta, SHAP, Relief, a Multi-Objective Evolutionary Algorithm (MOEA), and an Autoencoder. The feature rankings from each method are combined using a weighted aggregation strategy, followed by a stability check to retain only consistently important features. The final set of stable features was tested on the Wisconsin Diagnostic Breast Cancer (WDBC) dataset by training several state-of-the-art classifiers—ExtraTrees, CatBoost, LightGBM, XGBoost, and a soft Voting Classifier—within a 10-fold cross-validation framework. All classifiers achieved over 94% accuracy and ROC–AUC above 0.98, with ExtraTrees performing best at approximately 95.4% accuracy and 0.991 ROC–AUC. Compared with training on all original features, the proposed pipeline raises accuracy by roughly 4.5 percentage points (from 91% to 95.44 %) and improves mean ROC–AUC from 0.96 to 0.991.SHAP analysis highlights worst concave points, worst area, worst radius, mean concavity and area error as the most influential features—findings consistent with established clinical evidence. By combining probabilistic, evolutionary and neural-network–based selection techniques, the framework embodies key principles of soft computing and provides a reliable, interpretable basis for computer-aided diagnosis. It can be readily extended to larger or more heterogeneous datasets and to other imaging-based disease-detection tasks.

References

[1] Aličković E, Subasi A. Breast cancer diagnosis using GA feature selection and Rotation Forest. Neural Comput Appl 2017;28:753–63. https://doi.org/10.1007/s00521-015-2103-9.

[2] Al-Furat Al-Awsat Technical University Computer Center Administrator, Najaf, Iraq, .. AT, Department of System Programming, South Ural State University, Chelyabinsk 454080, Russia, .. H, Electronic Computer Center University of Diyala, Diyala, Iraq, .. AA, et al. A Comparative Analysis of Methods for Detecting and Diagnosing Breast Cancer Based on Data Mining. J Artif Intell Metaheuristics 2023;4:08–17. https://doi.org/10.54216/JAIM.040201.

[3] Rajpoot CS, Sharma G, Gupta P, Dadheech P, Yahya U, Aneja N. Feature Selection-based Machine Learning Comparative Analysis for Predicting Breast Cancer. Appl Artif Intell 2024;38:2340386. https://doi.org/10.1080/08839514.2024.2340386.

[4] Zhu J, Zhao Z, Yin B, Wu C, Yin C, Chen R, et al. An integrated approach of feature selection and machine learning for early detection of breast cancer. Sci Rep 2025;15:13015. https://doi.org/10.1038/s41598-025-97685-x.

[5] Fallahzadeh O, Dehghani-Bidgoli Z, Assarian M. Raman spectral feature selection using ant colony optimization for breast cancer diagnosis. Lasers Med Sci 2018;33:1799–806. https://doi.org/10.1007/s10103-018-2544-3.

[6] Yaqoob A, Verma NK, Mir MA, Tejani GG, Eisa NHB, Mamoun Hussien Osman H, et al. SGA-Driven feature selection and random forest classification for enhanced breast cancer diagnosis: A comparative study. Sci Rep 2025;15:10944. https://doi.org/10.1038/s41598-025-95786-1.

[7] Sumathi CP, Padmavathi MS. An experimental approach of applying boruta and elastic net for variable selection in classifying breast cancer datasets. Int J Knowl Eng Data Min 2019;6:356. https://doi.org/10.1504/IJKEDM.2019.105265.

[8] Liu J, Zhao L, Si C, Guan H, Dong X. Improved ReliefF-based feature selection algorithm for cancer histology. Biomed Signal Process Control 2023;85:104980. https://doi.org/10.1016/j.bspc.2023.104980.

[9] Shin J, Hassan N, Miah ASM, Suzuki T, Al Farhood S. Optimized Breast Cancer Classification Via SHAP-Based Feature Selection and Weighted Voting Ensemble Learning. IEEE Open J Comput Soc 2025:1–12. https://doi.org/10.1109/OJCS.2025.3610146.

[10] Shaon MdSH, Karim T, Shakil MdS, Hasan MdZ. A comparative study of machine learning models with LASSO and SHAP feature selection for breast cancer prediction. Healthc Anal 2024;6:100353. https://doi.org/10.1016/j.health.2024.100353.

[11] Yang F, Xu Z, Wang H, Sun L, Zhai M, Zhang J. A hybrid feature selection algorithm combining information gain and grouping particle swarm optimization for cancer diagnosis. PLOS ONE 2024;19:e0290332. https://doi.org/10.1371/journal.pone.0290332.

[12] Kazerani R. Improving Breast Cancer Diagnosis Accuracy by Particle Swarm Optimization Feature Selection. Int J Comput Intell Syst 2024;17:44. https://doi.org/10.1007/s44196-024-00428-5.

[13] Praveena V, Sujithra LR, Karthik S, Kavitha MS. Bio-Inspired ensemble feature selection and deep auto-encoder approach for rapid diagnosis of breast cancer. Multimed Syst 2023;29:3403–19. https://doi.org/10.1007/s00530-023-01168-w.

[14] Macías-García L, Martínez-Ballesteros M, Luna-Romera JM, García-Heredia JM, García-Gutiérrez J, Riquelme-Santos JC. Autoencoded DNA methylation data to predict breast cancer recurrence: Machine learning models and gene-weight significance. Artif Intell Med 2020;110:101976. https://doi.org/10.1016/j.artmed.2020.101976.

[15] Karuppasamy A, Zidoum H, Al-Rashdi MSS, Al-Bahri M. Optimizing breast cancer diagnosis with convolutional autoencoders: Enhanced performance through modified loss functions. Intell-Based Med 2025;11:100248. https://doi.org/10.1016/j.ibmed.2025.100248.

[16] Sowan B, Eshtay M, Dahal K, Qattous H, Zhang L. Hybrid PSO feature selection-based association classification approach for breast cancer detection. Neural Comput Appl 2023;35:5291–317. https://doi.org/10.1007/s00521-022-07950-7.

[17] Shaban WM. Insight into breast cancer detection: new hybrid feature selection method. Neural Comput Appl 2023;35:6831–53. https://doi.org/10.1007/s00521-022-08062-y.

[18] Halder RK, Lima MA, Uddin MN, Islam MdA, Saha A. Integrated feature selection-based stacking ensemble model using optimized hyperparameters to predict breast cancer with smart web application. Clin EHealth 2025;8:146–61. https://doi.org/10.1016/j.ceh.2025.08.001.

[19] Pagadala V, B V, Boinapalli S, Dhulipalla RK, Annapoorna S. A Quantum Convolutional Neural Network for Breast Cancer Classification using Boruta and GA-Based Feature Selection with Quantum Feature Maps. J Electron Electromed Eng Med Inform 2025;7:894–910. https://doi.org/10.35882/jeeemi.v7i3.792.

[20] Chhillar I, Singh A. An improved soft voting-based machine learning technique to detect breast cancer utilizing effective feature selection and SMOTE-ENN class balancing. Discov Artif Intell 2025;5:4. https://doi.org/10.1007/s44163-025-00224-w.