Feature Selection Using XGBoost on METABRIC Dataset for Survivability Breast Cancer Detection

Breast cancer is a highly lethal ailment that primarily affects women. The investigation of the survivability of breast cancer using different modalities remains an ongoing concern. One such approach involves analyzing the data obtained from The Molecular Taxonomy of Breast Cancer International Con- sortium (METABRIC). Regrettably, employing many modalities in the process results in an abundance of features in the data. We utilized the XGboost algorithm to carry out the process of feature selection by identifying the k-top features. We conducted a comparison of the findings using various machine learning clas- sifiers, including K-Nearest Neighbor, Support Vector Machine, Random Forest, and XGBoost. The results indicate that utilizing XGBoost as a feature selection method enhances the performance of all classifiers. The highest performing classifier is achieved by XGBoost and Random Forest, resulting in an accuracy value of 0.727273.