Breast cancer classification in mammographic images using fractal diffusion feature selection and k nearest neighbor

Breast cancer remains one of the leading causes of cancer-related mortality among women worldwide, highlighting the need for accurate and computationally efficient computer-aided diagnosis systems. Existing deep learning and optimization-assisted mammographic classification frameworks often suffer from high-dimensional feature redundancy, increased computational complexity, and limited interpretability, which reduce diagnostic reliability in clinical applications. To address these limitations, this study proposes a stochastic fractal search–based feature selection and classification framework for automated breast cancer diagnosis using mammographic images. The proposed framework integrates ROI-based handcrafted radiomic feature extraction, Z-score normalization, stochastic fractal diffusion optimization, and k-nearest neighbor (k-NN) classification within a patient-wise stratified 5-fold cross-validation strategy. The stochastic fractal search algorithm performs efficient wrapper-based feature optimization by eliminating redundant diagnostic attributes and identifying the most discriminative feature subset from CBIS-DDSM and INbreast datasets. Experimental results demonstrate that the proposed SFS-kNN framework achieved classification accuracies of 97.3% and 98.7%, precision values of 97.1% and 98.5%, recall values of 97.5% and 98.5%, F1-scores of 97.3% and 98.6%, and AUC values of 0.991 and 0.995 on the CBIS-DDSM and INbreast datasets, respectively. In addition, the proposed framework reduced the original 30-dimensional feature space to an average of 12.67 optimized features, thereby significantly improving computational efficiency and convergence stability. Comparative analysis further confirms that the proposed SFS-kNN framework outperforms conventional optimization-based classifiers in terms of diagnostic accuracy, feature reduction capability, robustness, and interpretability, demonstrating its effectiveness as a reliable lightweight clinical decision support framework for breast cancer classification.