Location
Hilton Hawaiian Village, Honolulu, Hawaii
Event Website
https://hicss.hawaii.edu/
Start Date
3-1-2024 12:00 AM
End Date
6-1-2024 12:00 AM
Description
Learning Analytics (LA) has been leveraged as a tool to analyze and improve educational processes by informing its stakeholders. LA for student profiling focuses on discovering learning patterns and trends based on diverse features extracted from trace data. Prior studies have used classical clustering methods to group students and understand the study patterns of each cluster. However, variations within the clusters are still large making it difficult to draw concrete insights into the relation between study behaviors and learning outcomes. In this work, we leverage anomaly detection and eXplainable AI techniques to distinguish between normal and abnormal study patterns and to possibly discover unexpected patterns that are not apparent from clustering alone. We perform external validation to check the generalizability and compare the insights on study patterns from our method to be at par with insights gained from previous studies.
Recommended Citation
Tiukhova, Elena; Vemuri, Pavani; Óskarsdóttir, María; Poelmans, Stephan; Baesens, Bart; and Snoeck, Monique, "Discovering Unusual Study Patterns Using Anomaly Detection and XAI" (2024). Hawaii International Conference on System Sciences 2024 (HICSS-57). 3.
https://aisel.aisnet.org/hicss-57/da/learning_analytics/3
Discovering Unusual Study Patterns Using Anomaly Detection and XAI
Hilton Hawaiian Village, Honolulu, Hawaii
Learning Analytics (LA) has been leveraged as a tool to analyze and improve educational processes by informing its stakeholders. LA for student profiling focuses on discovering learning patterns and trends based on diverse features extracted from trace data. Prior studies have used classical clustering methods to group students and understand the study patterns of each cluster. However, variations within the clusters are still large making it difficult to draw concrete insights into the relation between study behaviors and learning outcomes. In this work, we leverage anomaly detection and eXplainable AI techniques to distinguish between normal and abnormal study patterns and to possibly discover unexpected patterns that are not apparent from clustering alone. We perform external validation to check the generalizability and compare the insights on study patterns from our method to be at par with insights gained from previous studies.
https://aisel.aisnet.org/hicss-57/da/learning_analytics/3