Machine Learning Approach on Breast-Cancer Prediction with Smote Technique

Breast cancer is a notable global concern in female health due to its rapid spread and high mortality rate. Early detection is crucial, albeit challenging, in combating this disease. Machine learning and ensemble techniques have been pivotal in advancing early detection, showing substantial growth in predictive accuracy. This study presents a robust approach for accurately predicting breast cancer by employing synthetic minority oversampling techniques (SMOTE) alongside recursive feature elimination for optimal feature selection. We compare the performance of various algorithms —including Logistic Regression, Random Forest, Support Vector Machine, and Gradient Boosting—focusing on hyperparameter optimization via grid search cross-validation to enhance case performance. Ensemble techniques, particularly voting classifiers, achieved exceptional accuracy, yielding 99.55% and 99.46% on the WBCD and WBC test datasets, respectively, and surpassed individual algorithms across metrics such as precision, recall, and F1 score. This research underscores the vital role of ensemble methods and feature selection in advancing breast cancer detection, providing a highly accurate, reliable, and efficient diagnostic tool that sets a new standard in biomedical prediction.