Comparison of ANN Backpropagation Algorithm and Random Forest Regression in Predicting the Number of New Students

Abstract

Higher education institutions are educational units located at a higher level after high school or vocational school. Catholic University Darma Cendika Surabaya (UKDC) faces challenges in managing the admission of new students due to variations in the number of prospective students applying to each department, which is also influenced by changing trends in interests and job needs in Indonesia. The use of Artificial Neural Network with Backpropagation and Random Forest Regression algorithms for comparing the prediction of new student admissions in the following year will be beneficial for the administration of Catholic University Darma Cendika Surabaya (UKDC) to gain a clearer understanding of the dynamics of admissions and to support decision making in the future development of the university. The predicted number of students joining Catholic University Darma Cendika Surabaya (UKDC) in the 2024 period using Artificial Neural Network is 219 students with a Mean Squared Error (MSE) of 0,1046 and Root Mean Square Error (RMSE) of 0,32.

Keywords

• artificial neural network, backpropagation
• random forest regression
• msermse