Design and Development of an Efficient Risk Prediction Model for Cervical Cancer

Abstract

Cervical cancer is a major public health concern, especially in low- and middle-income countries. Lifestyle choices to some extent have an effect on causing cervical cancer. Most cervical cancers are caused by the sexually transmitted infection caused by the Human Papillomavirus (HPV). However, only persistent HPV infections lead to progression to pre-cancer and cancer. The persistence of this infection is influenced by many factors namely, age, sexually transmitted infections, number of sexual partners, age at first sexual intercourse, number of deliveries, tobacco consumption, etc. Risk-based prediction algorithms help to stratify women with a high risk to develop cervical cancer and screen them on a priority basis. In this study, a model has been developed to predict the risk of cervical cancer based on one’s lifestyle choices. Important features have been delineated using the Extreme Gradient Boosting (XGBoost) Classifier. After oversampling, the data is fed into the model for training and testing. The Gradient Boost model was chosen to arrive at an accurracy of 98.9%. This model can be effective to associate risk factors with cervical cancer prediction which can help the in the effective prevention and management of cervical cancer.

Keywords

• Machine learning
• digital health
• cervical cancer
• human papillomavirus
• risk factors
• predictive modeling