Early Breast Cancer Detection in Coimbra Dataset Using Supervised Machine Learning (XGBoost)

Abstract

Worldwide, breast cancer (BC) represents one of the serious health concerns for adult females. The early detection and accurate prediction of risks are vital for the provision of optimum care and enhancement of patient outcomes. In the past few years, promising large data merging and ensemble learning algorithms appeared for the purpose of classification and prediction of BC risk. In the area of medical applications, methods of machine learning (ML) are crucial. Early diagnosis is necessary for a more efficient carcinoma treatment. This study’s aim is to classify the carcinoma with the use of the 10 predictors that are found in Breast Cancer Coimbra dataset (BCCD). Presently, early diagnoses are necessary. The rates of cancer survival could be raised in the case where it is discovered early. Methods of machine learning offer effective way for data classifying and making early disease diagnoses. This study utilizes BCCD for the classification of BC cases utilizing XGBoost algorithm. Based on performance criteria, early detection of BC is the primary goal. The XGBoost classifier in this research achieved 98% precision, 98.32% accuracy, 99% f1-score, and 97% recall.

Keywords

• machine learning
• xgboost
• z-score
• bccd