An XGBoost-Based Classification Method to Classify Breast Cancer
Kishwar Sadaf, Jabeen Sultana, Nazia Ahmad · Auerbach Publications eBooks · 2023
The people affected with this dreadful disease known as breast cancer are susceptible to complex health issues and the mortality rate has increased day by day. Diagnosing breast cancer is quite a laborious task and involves careful observation to screen the breast to identify irregularities in the breast because of many underlying features. If breast cancer is diagnosed at early stages, then it will assist the doctors and patients to handle the disease with ease. Therefore, it is worth classifying breast cancer disease to identify which stage of breast cancer the person is currently belonging to. This led to the emergence of using machine learning techniques to classify and diagnose breast cancer data. Currently, classifiers of machine learning are used in wide areas of research to perform classification, clustering and mining the rules accordingly. In this chapter breast cancer dataset has been collected from the UCI machine learning repository, which has 569 instances with 31 attributes. In this chapter, we propose a classification of breast cancer data into benign and malignant classes using XGBoost. We use XGBoost in an intuitive way where we extract the features of the dataset which are considered important by it i.e., XGBoost being used as a feature selection method. We use these features in different classification schemes. Testing is carried out to check the accuracy of the model built by the classifier. Result analysis reveals that important features yielded by XGBoost significantly increase the model accuracy of different classification methods.