TY - GEN
T1 - Detection of Student Modelling Anomalies
AU - Sosnovsky, Sergey
AU - Müter, Laurens
AU - Valkenier, Marc
AU - Brinkhuis, Matthieu
AU - Hofman, Abe
PY - 2018/1/1
Y1 - 2018/1/1
N2 - As the modern TEL tools gain wider adoption in real educational contexts, they start facing important practical problems. One such problem for adaptive educational systems is the reliability of their student modelling mechanisms. Even when such a mechanism has been tested and calibrated to represent students’ knowledge reasonably well, the student herself can become a source of problems. Students can use the system in a non-intended way, exhibit long periods of off task behaviour, try gaming the system, seek help of parents or peers, etc. Such usage patterns will manifest themselves in sequences of activity that do not represent student abilities and will result in student modelling anomalies causing subsequent suboptimal adaptive interventions from the system. This would be very important for a system that is used in real classrooms with younger children, especially, when it is also available at home as a supporting tool for independent work. This paper reports a study of such a system – Math Garden. Several user modelling anomalies have been detected in its logs. First steps towards building an automated tool for on-the-fly student modelling anomaly detection are reported.
AB - As the modern TEL tools gain wider adoption in real educational contexts, they start facing important practical problems. One such problem for adaptive educational systems is the reliability of their student modelling mechanisms. Even when such a mechanism has been tested and calibrated to represent students’ knowledge reasonably well, the student herself can become a source of problems. Students can use the system in a non-intended way, exhibit long periods of off task behaviour, try gaming the system, seek help of parents or peers, etc. Such usage patterns will manifest themselves in sequences of activity that do not represent student abilities and will result in student modelling anomalies causing subsequent suboptimal adaptive interventions from the system. This would be very important for a system that is used in real classrooms with younger children, especially, when it is also available at home as a supporting tool for independent work. This paper reports a study of such a system – Math Garden. Several user modelling anomalies have been detected in its logs. First steps towards building an automated tool for on-the-fly student modelling anomaly detection are reported.
KW - Adaptive educational system
KW - Educational data mining
KW - Student modelling
KW - Student modelling anomaly
UR - http://www.scopus.com/inward/record.url?scp=85053193139&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-98572-5_41
DO - 10.1007/978-3-319-98572-5_41
M3 - Conference contribution
AN - SCOPUS:85053193139
SN - 9783319985718
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 531
EP - 536
BT - Lifelong Technology-Enhanced Learning - 13th European Conference on Technology Enhanced Learning, EC-TEL 2018, Proceedings
PB - Springer
T2 - 13th European Conference on Technology Enhanced Learning, EC-TEL 2018
Y2 - 3 September 2018 through 6 September 2018
ER -