A Computer-Aided Risk Assessment and Major Factors Analysis System for Cervical Cancer

Cervical cancer stands as a primary cause of premature death among women worldwide with over 85% of fatalities in developing nations. This cancer is influenced by various risk factors, and early detection and treatment overseen by expert medical professionals are crucial to minimizing its negative effects. This study extensively investigates demographic and lifestyle data to identify important factors associated with cervical cancer. This work aims to investigate the feature importance of habits and practices associated with risks of cervical cancer using UCI cervical cancer risk factors dataset. This investigation gives the idea of essential risk factors to enhance the understanding of vaginal health. By employing feature importance algorithm, we thoroughly examine a wide range of variables including demographic, behavioral, and clinical attributes. Robust statistical techniques have enabled the identification of prominent risk factors. Additionally, a computer-aided assessment system has been developed to predict the risk of cervical cancer using machine learning framework. This system simplifies the process, saving time, resources, and reducing complications associated with manual prediction or delayed diagnosis. Six classifiers are evaluated for the proposed predictive model, with random forest consistently outperforming (highest accuracy of 99.36%) than other classification approaches. The future direction of this work includes multi-stage severity assessment and translation of the predictive cervical cancer risk assessment model into mobile or web-based applications.