Abstract

Breast cancer poses a significant challenge, necessitating robust predictive models for early diagnosis and treatment planning. This study aims to enhance a machine learning model using Adaptive Synthetic Sampling (ADASYN) to address class imbalance in the Breast Cancer Coimbra Dataset (BCCD). ADASYN generates synthetic samples for the minority class, balancing the dataset and improving model sensitivity to malignant cases. We employed the Light Gradient Boosting Machine (LightGBM) model, known for its high efficiency, and Hyperparameter tuning using Grid Search with cross-validation optimized LightGBM. Our findings show a substantial improvement in performance, with the optimized model significantly outperforming existing approaches. The model achieved higher accuracy, precision, recall, and F1-score, demonstrating the effectiveness of addressing class imbalance and hyperparameter optimization. Enhanced predictive accuracy can lead to earlier detection and more precise treatment planning, ultimately improving patient outcomes. Future research will explore advancements in ensemble methods and deep learning architectures.