Machine learning for assisting cervical cancer diagnosis: An ensemble approach

Abstract Cervical cancer remains one of the most prevalent gynecologic malignancies, worldwide. As cervical cancer is a highly preventable disease, therefore, early screening represents the most effective strategy to minimize the global burden of cervical cancer. However, due to scarce awareness, lack of access to medical centers, and highly expensive procedures in developing countries, the vulnerable patient populations cannot afford to undergo examination regularly. A novel ensemble approach is presented in this paper to predict the risk of cervical cancer. By adopting a voting strategy, this method addresses the challenges associated with previous studies on cervical cancer. A data correction mechanism is proposed to improve the performance of the prediction. A gene-assistance module is also included as an optional strategy to enhance the robustness of the prediction. Multiple measurements are performed to evaluate the proposed method. The results indicate that the likelihood of developing cervical cancer can be effectively predicted using the voting strategy. Compared with other methods, the proposed method is more scalable and practical.