Optimizing Machine Learning for Breast Cancer Detection by Hybrid Metaheuristic Approach

Breast cancer ranks as one of the top causes of cancer-related deaths among women worldwide, highlighting the need for improved diagnostic methodologies. Machine learning offers promising solutions; however, the models' performance highly relies on optimal hyperparameter configurations. The goal of this research is to improve the precision of machine learning models in diagnosing breast cancer by employing a novel optimization strategy that integrates the metaheuristic algorithm, namely, the Sine Cosine Algorithm with the Opposition-Based Learning mechanism for hyperparameter tuning. We applied five machine learning classifiers: K-Nearest Neighbors, Decision Trees, Random Forest, Logistic Regression, and Support Vector Machines to the Breast Cancer Wisconsin dataset. The proposed optimization method was utilized to fine-tune the hyperparameters of these classifiers to improve the performance. Model performance was rigorously evaluated using fivefold cross-validation approach and it was compared to existing methods. The optimization process led to significant improvements in model performance, with the enhanced classifiers outperforming their baseline configurations and existing works. The findings highlight the success of the suggested method in navigating the hyperparameter search space, leading to more accurate and reliable breast cancer diagnoses. Integrating the Sine Cosine Algorithm with Opposition-Based Learning for hyperparameter tuning presents a powerful approach to refining the machine learning classifiers for breast cancer diagnosis. This study underscores the potential of advanced optimization methods in improving the diagnostic capabilities of machine learning models, contributing to timely diagnosis and treatment of breast cancer.