Empirical Evaluation of Breast Cancer Detection Using Machine Learning and Deep Learning Techniques

Breast cancer is a prevalent and serious condition affecting women worldwide. Early and accurate diagnosis is critical for determining cancer prognosis and improving patient outcomes. Despite advancements in diagnostic technologies, existing methods often lack accuracy and generalizability, highlighting the need for reliable early detection and classification systems. This study preprocesses breast cancer datasets to ensure clean data, extracts key features, and selects them to enhance model performance. The datasets are divided into training and testing sets, where the training data builds the model, and the testing data validates its accuracy. Cases are classified as normal or abnormal, with abnormal cases further categorized as benign or malignant. Performance is evaluated using metrics such as accuracy, precision, and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{F 1}$</tex>-score. The results demonstrate promising accuracies of 97.81 % with 70 % training data, 98 % with 80 %, and an ideal 100 % with 90 % training data. These findings underscore the effectiveness of the proposed system in improving breast cancer detection and classification, emphasizing the transformative role of artificial intelligence in revolutionizing breast cancer diagnosis and patient care.