Naslov
Predicting Students’ Academic Performance Based on Enrolment Data

Autori
Zacharias, Dermatis ; Athanasios, Anastasiou

Izvornik
International Journal of Innovation and Economic Development (1849-7020) 6 (2020), 3; 36-45

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
Neural networks, Educational Data mining, Student’s academic performance

Sažetak
Efficient education is key to the development and progress of modern society. Identifying factors that affect students’ academic performance is a very important step towards efficient education. With fast IT development and lower prices, universities start to collect a huge amount of data. With data mining methods and techniques, universities can use this data, analyze it and get hidden and useful information. This paper presents a model for predicting students’ academic performance based on enrolment data using one of the data mining techniques, Neural network. The enrolment data consists of demographic and economic data and information about previous education. Students’ academic performance is measured by grade point average in university, and based on that, students are divided into two groups. One group consists of students with a grade point average below 3.5, and the other group consists of students with a grade point average above 3.5. This model may represent the first step for educators to early intervene and reduce the percentage of students leaving universities. They could offer students who are classified below average some additional classes to overcome the more difficult courses because of insufficient prior knowledge, thereby, increasing their likelihood of continuing their studies.

Izvorni jezik
Engleski

Znanstvena područja
Informacijske i komunikacijske znanosti

POVEZANOST RADA