Prediction model for cervical cancer by using machine learning

Abstract

The aim of this study is to compare predictive performance of cervical cancer prediction models using machine learning techniques, including decision trees, gradient boosted trees, random forests, and deep learning. The study collected data from Kaggle.com, comprising 110 data points, with 21 independent variables, such as age, number of sexual partners, age of first sexual intercourse, number of pregnancies, smoking per year, contraceptive pill use, year of contraceptive pill use, intrauterine device use, years of intrauterine device use, sexually transmitted infections, genital herpes, genital warts, syphilis, cervical infection, herpes simplex virus infection, rice-sized cervical polyps, AIDS, HIV infection, HPV infection, and hepatitis B virus infection. The dependent variable is binary, classifying individuals into two groups: those with cervical cancer and those without. The study results indicate that the deep learning-based cervical cancer prediction model performs with the highest predictive efficiency, achieving an accuracy rate of 95.45%.