Breast Cancer Modeling and Prediction Combining Machine Learning and Artificial Neural Network Approaches

Breast cancer strikes women more frequently than it strikes males, which is a major factor in the rising mortality rate for women. Breast cancer, which predominantly concerns women (roughly 1% of cases include non-females), will impact one in eight women in their lifetime. Breast carcinoma, one of the deadliest malignancies, is the main cause of cancer-related deaths in women. Today, early diagnosis and prognosis are crucial to increase survival rates and finally bring them down. Cancer researchers have a number of difficulties when attempting to differentiate between benign and malignant tumors as well as when attempting to make judgments about mild and metastatic breast cancer. Research has turned its attention to machine learning techniques, which have been shown to be successful in the early identification and prognostication of breast cancer. In proposed article, we utilized eight machine learning techniques to the Breast Cancer Wisconsin Diagnostic dataset: Support Vector Machine, Logistic Regression, XGBoost, CatBoost, Random Forest, Artificial Neural Network, Decision Tree, K-Nearest Neighbours. After getting the data, a performance assessment and comparison of these different classifiers is done. The foremost goal of this study is to detect and utilize machine learning to detect breast cancer and determine which method is more efficient in terms of confusion matrices, accuracy, and precision. The Support Vector Machine and Artificial Neural Network have demonstrated superior performance than all other classifiers, with an accurate result of 98.08%.