Boosting vs. Traditional Algorithms: Unveiling the Best Approach for Breast Cancer Prediction

Breast cancer is still a major health problem worldwide and requires accurate and efficient detection methods to improve patient outcomes. This study compares traditional machine learning models such as K-nearest neighbor, support vector machine, and MLP neural network with boosting algorithms such as histogram-based gradient boosting, CatBoost, and LightGBM in breast cancer classification. We evaluated each model based on precision, precision, recall, F1 score, and area under the curve (AUC) using the UCI breast cancer dataset. The traditional model shows good performance, and KNN and SVM achieve balanced precision and recall. Histogram-based gradient boosting was found to be the best performing boosting model, achieving an AUC of 0.9970 and showing good discriminative power. Comparative analysis of confusion matrices and ROC curves reveals that histogram-based gradient boosting and neural network MLP are the most reliable models for accurate classification. These findings highlight the potential of machine learning in improving breast cancer diagnosis and provide insight into model selection based on specific clinical needs.Future research may consider hybrid approaches or ensemble techniques to further improve classification performance in different patient populations.