Long-term prediction of master student influx in programs Artificial Intelligence and Applied Data Science

Abstract

The aim of this paper is to predict the long-term influx of new master students in the Artificial Intelligence program at Utrecht University. A VARMA model was trained on 81.25% of the data, the remaining 18.75% of the data was used as test set. The VARMA model required at least two time series data sets which influence each other. The first was the past influx numbers of students to the Artificial Intelligence program. The second was the number of bachelor diplomas for Artificial Intelligence. The predictions show that the number of new students will rise from 119 in 2022 to 194 in 2031. These results were predicted with a MAPE value of 6.75, which means the model is a good fit. Furthermore the results show a linear trend in the increase of new master students.