Implementation of Machine Learning Algorithms for Early Detection of Cervical Cancer Based on Behavioral Determinants

Abstract

Cervical cancer is a disease that affects women and has the highest mortality rate after breast cancer. Early detection of cervical cancer is critical at this time, so cervical cancer patients are decreasing. Many women, especially in Indonesia, are less concerned about the dangers of cervical cancer, even though if detected earlier, this disease will be easier to treat. One alternative for early detection can use machine learning algorithms. The machine learning algorithms used in this study are Naïve Bayes (NB), Logistic Regression (LR), Decision Tree (DT), SVM, and Random Forest. In this study, a random under-sampling method was employed, which had no uses in any prior research. This technique makes the accuracy of the five algorithms even better. The research results show that NB has an accuracy rate of 91.67%, LR has an accuracy rate of 87.5%, DT has an accuracy rate of 81.81%, SVM has an accuracy rate of 75%, and RF has the highest accuracy rate of 94.45%. This research shows that the best model is RF or Random Forest