Ensemble-Based Machine Learning Models for the Preemptive Diagnosis of Cervical Cancer using Clinical Data

Early detection and classification of a cancer type can aid the patient's subsequent clinical therapy. Cervical cancer is a prevalent type of cancer that affects a woman's cervix because of abnormal cell growth. It is the fourth most common disease in women globally, and early detection has the potential to save lives. In spite of the fact that previous studies achieved outstanding results, their models were highly computationally complex due to the significant number of features incorporated. Therefore, this study is intended to use clinical data to diagnose cervical cancer in its early stages and investigate the effect of Recursive Feature Elimination (RFE) in reducing the number of attributes. The models were trained and tested using a dataset obtained from the Machine Learning Repository at the University of California at Irvine (UCI). The models included Extreme Gradient Boosting (XGBoost), Adaptive Boosting (AdaBoost), and Extra Trees (ET). The findings showed that XGBoost and AdaBoost performed the best, with an accuracy rate of 99.79%, recall rate of 100%, and precision rate of 99.6%, using only 8 features.