Ovarian Cancer Detection Using SMOTE-XGBoost Classifier Model

Abstract

Ovarian cancer diseases are considered to be among the most fatal malignancies affecting women. In this study, the power of XGBoost and SMOTE are combined to provide a novel method for increasing classification accuracy in unbalanced datasets. The purpose of the study's classification challenge is to equalise the representation of majority and minority classes in the dataset. By creating artificial instances of the minority class, to improve the dataset's balance by addressing class imbalance through the application of the SMOTE algorithm. The process of feature selection makes use of the strong feature importance scores that XGBoost offers, making it possible to identify important variables for classification. The XGBoost model achieves overall accuracy of 94.14%, which is a considerable improvement in classification accuracy. This work is an important addition to the fields of data science and machine learning because it demonstrates how class imbalance may be efficiently addressed and model performance can be enhanced by machine learning techniques.