A Soft Computing-Based Intelligent Framework for Enhanced Breast Cancer Detection

Breast cancer continues to be a critical health concern globally, and patient prognoses are markedly improved with early identification. Deep learning techniques have emerged as excellent instruments for the analysis of medical pictures, and their application in breast cancer detection has the potential to enhance analytical accuracy. This research introduces a unique methodology called mutual information filter force-driven NeuroConvolveX (MIFF-NCX) for feature selection aimed at improving early-stage breast cancer identification. Proposed methodologies integrate the advantages of deep learning in extracting intricate features from medical datasets with feature selection techniques that emphasize the most valuable attributes. We utilize the original Breast Cancer dataset to assess proposed methodologies and data analysis. In terms of measures like F-score, sensitivity, ROC curve, specificity, and accuracy curves, the presentation of the suggested models is assessed and contrasted with existing techniques. The findings of this study show that recommended techniques significantly increase the efficacy and efficiency of deep-learning models for identifying breast cancer in its early stages.