Late payment tuition fees predicted with 58.70% accuracy using Random Forest-ROS.
The article compares two methods, Random Forest and AdaBoost, to predict late payment of tuition fees. They used different techniques to balance the data and found that Random Forest with Random Oversampling had better predictions than AdaBoost with Random Undersampling. The important factors for predicting late payments were the household's electric capacity, father's income, and number of children in the family.
