From Diagnosis to Prognosis: Enhancing Breast Cancer Survival Predictions Through the Application of Machine Learning and Feature Selection Techniques

Breast cancer remains a significant global health challenge, necessitating the development of advanced diagnostic methods. This study explores the use of machine learning algorithms to predict breast cancer diagnoses and survival outcomes based on clinical data. By applying preprocessing techniques and using publicly available datasets, the research evaluates multiple classification models, including KNN, Logistic Regression, SVM, Naive Bayes, Decision Trees, and ensemble methods like Random Forest, XGBoost, and LightGBM. The models were evaluated for their performance in terms of accuracy, precision, and recall. Results show that LightGBM achieved the highest accuracy, demonstrating the potential of machine learning in enhancing early breast cancer detection. This paper highlights the promising role of AI-driven approaches in improving medical decision-making, particularly in resource-constrained settings, and suggests areas for future enhancement, including handling class imbalance and incorporating more extensive real-world data.