Аңдатпа
Breast cancer remains a leading cause of mortality among women, and early detection is crucial for improving patient outcomes. However, current detection methods, including mammography and BMI-based approaches, have limitations in terms of accuracy and reliability. This research proposes an innovative breast cancer detection using the Breast Cancer Detection and Risk Assessment (BCDRA) system. BCDRA integrates the features of convolutional neural networks (CNNs) and transfer learning, with an optimized feature selection process designed specifically for mammographic images and patient clinical data. The system follows a multi-stage pipeline: Data preprocessing and augmentation, Automated feature extraction using CNNs, Transfer learning for enhancing model accuracy, and Risk prediction using a hybrid classification-regression algorithm. This pipeline is supported by the Breast Cancer Prediction Algorithm (BCPA), which identifies early-stage anomalies and provides real-time risk assessment based on personalized risk factors such as age, family history, and genetic markers. By integrating a real-time risk assessment module and offering higher diagnostic precision, BCDRA aims to revolutionize breast cancer screening, allowing for earlier interventions and improved patient care, ultimately contributing to better survival rates. BCDRA implemented using both Python and $\mathbf{R}$ programming environments and validated across multiple public and clinical datasets, demonstrating significant improvements in accuracy, sensitivity, and specificity over traditional detection methods.