UOC tests AI system to flag up students at risk of failing courses

A Universitat Oberta de Catalunya (UOC) research team headed up by David Bañeres from the Systems, Software and Models (SOM) Research Lab, which is affiliated to the Internet Interdisciplinary Institute (IN3), has developed a tool based on artificial intelligence (AI) techniques for identifying students at risk of failing their courses. This is done by analysing large data sets and applying algorithms that provide predictive models for student performance.

This study, recently published in the US open access journal Applied Sciences, has been carried out by Bañeres with the support of the UOC's online learning innovation, transformation and transfer hub, the eLearn Center.

Over the course of the past five years, the UOC has been collecting large amounts of anonymized data regarding its students, their on-campus activity and academic results in its data mart system, which has made it possible to analyse and identify student behaviour patterns that would be undetectable without the use of technology.

At-risk student alerts

Bañeres and his team are working on the project titled New Goals: Learning Intelligent System (LIS) to develop an early warning system for students at risk of failing a course. In the words of the researcher, "The system uses an artificial intelligence prediction model which is trained via the analysis and independent processing of historical data from each course. This generates a predictive model based on the patterns identified, which gives us information on how the students enrolled in the courses might perform."

The information generated by the system is used to assign each student a colour based on a traffic light system: red if they are at risk of failing; orange if the system cannot ensure that they will pass the subject and green if the model indicates that the student will pass. The relevant teaching staff then use the results of the analysis carried out by the system to send students a personalized message containing information on their level of risk, thereby establishing a basis by which students can work on improving their academic performance.

According to Bañeres, "One of the main concerns about using artificial intelligence in education centres is its potential for replacing teachers. This project draws on the expertise of the teacher as a necessary requirement. It is a teaching support system and is not intended to replace anyone."

Real-time application for 2022

Work on the project began in February 2019, with three pilot tests conducted to date involving the participation of almost 3,000 students studying on various courses within the Faculty of Computer Science, Multimedia and Telecommunications; the Faculty of Economics and Business, and the Faculty of Law and Political Science.

The results of the first pilot tests have shown that the accuracy of the predictions made by the system increase according to the variety and quantity of the data supplied. The rate of accuracy in predicting whether a student may have problems passing a course at the start of a semester, when little student information is available, is close to 60% but by halfway through the semester, that value increases to almost 90%. The pilot test currently under way is looking at performing real-time predictions in relation to a student's likelihood of dropping out of a course, as well as other factors, such as whether they will submit assessment activities, while facilitating personalized real-time interventions.

Bañeres described the response from students as positive, "in particular with regard to the personalized messages sent to them by their professors and the fact that the new system enables them to work together with the staff to optimize their academic performance".

In the coming months, the team plans to conduct further pilot tests on courses run by the University's other faculties. The project is scheduled to be completed in February 2022, by which time the team hope to have honed the system into the most accurate detection model possible that can be deployed in real time, i.e. to assist students with their daily tasks.

Artificial intelligence is emerging as a key tool to provide educational support and facilitate optimal learning tailored to each individual student. One of the benefits of this type of technology is that it permits the provision of student support and monitoring services on a grand scale.

Since its foundation 25 years ago, the UOC has distinguished itself by being a unique institution that places its student at the centre of their learning process with the support of the most innovative technologies. By opting to use artificial intelligence, which represents one of the most potentially powerful technologies for application in the world of education – as well as in health and other areas – the UOC aims to further personalize and optimize the learning process of its students.

This new tool, which is expected to be able to be comprehensively applied across all UOC faculties as of 2022, has the potential to be transferred to other higher education institutions.

This research promotes Sustainable Development Goal (SDG) 4 – Quality education.

 

Reference article

David Bañeres, M. Elena Rodríguez, Ana Elena Guerrero-Roldán & Abdulkadir Karadeniz (2020). "An Early Warning System to Detect At-Risk Students in Online Higher Education". Applied Sciences, 10(13), 4427. DOI:  https://doi.org/10.3390/app10134427.

UOC R&I 

The UOC's research and innovation (R&I) are helping 21st-century global societies to overcome pressing challenges by studying the interactions between ICT and human activity, with a specific focus on e-learning and e-health. Over 400 researchers and 50 research groups work among the University's seven faculties and two research centres: the Internet Interdisciplinary Institute (IN3) and the eHealth Center (eHC).

The United Nations' 2030 Agenda for Sustainable Development and open knowledge serve as strategic pillars for the UOC's teaching, research and innovation. More information:research.uoc.edu. #UOC25years