Machine Learning in Breast Cancer Diagnosis: Model Comparison and Practical Implications

Early and accurate diagnosis of breast cancer can significantly increase the chances of successful treatment, and modern digital health systems (eHealth) are increasingly using machine learning models to support doctors in decision-making. This paper analyzes and compares different machine learning models–including XGBoost, logistic regression, K-Nearest Neighbors (KNN), Naive Bayes, Decision Tree and Random Forest–to determine which model most effectively distinguishes benign from malignant tumors based on diagnostic data. The models were tested and evaluated according to accuracy, precision, sensitivity, and F1-score. At the same time, data were analyzed in detail through visualizations, including correlation of medical parameters, distribution of feature values, and display of confusion matrix. The contribution of this work is reflected in a comprehensive comparative analysis of several machine learning models, providing detailed insight into their performance and application possibilities in real clinical environments. In addition, the paper highlights the importance of model interpretability. It analyzes key challenges in their practical application, including access to quality data, balancing accuracy and sensitivity, and integration into existing eHealth systems. This study shows how different approaches can be used in medical diagnostics, highlighting the potential of machine learning in creating reliable systems for the early detection of breast cancer and supporting medical workers in their daily work.