Development of a Machine Learning-Based Diagnostic Support System for Early Identification of Breast Cancer Using Mammography Numerical Data

Breast cancer remains a major global health challenge, particularly in regions lacking consistent screening and early intervention. Accurate and timely diagnosis is vital for reducing mortality and improving treatment options, but traditional methods often require resources and expertise unavailable in many settings. This study develops and evaluates a machine learning-based diagnostic support system for early breast cancer detection using mammography numerical data. The dataset includes rich clinical and radiological features, allowing for an interpretable and scalable diagnosis approach. Preprocessing addressed outliers and class imbalance using the Synthetic Minority Oversampling Technique (SMOTE). The Random Forest classifier achieved 96% accuracy in predicting malignancy, demonstrating strong performance across internal and external validations. Comparative analysis showed that Random Forest outperformed other algorithms, including K-Nearest Neighbors, Support Vector Machine, and Logistic Regression, particularly in handling structured medical data and imbalanced classes. By automating the analysis of key numerical parameters from mammography, the system provides clinicians with rapid and reliable decision support, which is especially valuable in resource-limited environments. Importantly, early detection made possible by this AI-driven system may result in better case management and improved patient outcomes, ultimately reducing the societal and economic toll of breast cancer By facilitating reliable early detection, this system may support clinicians in prioritizing cases and tailoring treatments, ultimately improving patient outcomes and reducing the burden of breast cancer.