A Comparative Study of Machine Learning Algorithms for Breast Cancer Classification

Breast cancer remains a global health challenge, demanding effective early detection and classification methods for improved patient outcomes. The integration of machine learning techniques has emerged as a promising approach, aiding healthcare professionals in the complex task of breast cancer diagnosis. This study places a particular focus on the pivotal roles of manual feature selection and outlier handling within the domain of machine learning-based breast cancer classification. Accuracy serves as the primary metric for assessing the performance of the models. Results unveil the substantial impact of the feature selection process, showcasing outstanding accuracy rates. Most notably, Random Forest model demonstrates exceptional proficiency, achieving an impressive 96% accuracy alongside a commendable F1 score of 0.93 after feature selection. Additionally, Logistic Regression, Support Vector Machine, and Decision Tree models exhibit robustness in their capacity to distinguish between breast cancer cases, underlining the instrumental role of the selected features in classification. This study not only advances our understanding of breast cancer classification but also beckons future research endeavors. The exploration of more extensive datasets and alternative machine learning algorithms stands as a promising avenue to further elevate classification performance. Such advancements promise more accurate breast cancer diagnoses, ultimately leading to enhanced outcomes of patient care.