@inproceedings{56a6f87d3dd14ecd9ad612966fe27b05,
title = "Analyzing Student Behavior from Moodle Data Using Process Mining and Deep Learning",
abstract = "This research examines the amalgamation of process mining and deep learning to evaluate student interaction behavior on the Moodle learning platform. We used data from the CHEM262 course at Jordan University of Science and Technology (JUST) to build instance-level graphs that show what each student did using the Building Instance Graph (BIG) algorithm. After that, these graphs were used to train a Multi-Layer EdgeConv neural network, which was able to tell the difference between successful and struggling students with 99.08\% accuracy. The process mining analysis additionally indicated that sustained engagement and regular grade monitoring were significant predictors of academic achievement. These insights show how combining process mining with graph-based deep learning models could help find students who are at risk and help teachers make timely, data-driven decisions about how to help them.",
keywords = "Deep Learning, EdgeConv, Educational Data Mining, Moodle, Process Mining",
author = "Agolah, \{Rami Abu\} and Malak Abdullah and Alex Mircoli and Mahmoud Al-Ayyoub and Dana Elrushaidat",
note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 2025 International Conference on Cybersecurity and AI-Based Systems, Cyber-AI 2025 ; Conference date: 01-09-2025 Through 04-09-2025",
year = "2025",
doi = "10.1109/Cyber-AI66431.2025.11233793",
language = "English",
series = "2025 International Conference on Cybersecurity and AI-Based Systems, Cyber-AI 2025",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "143--148",
editor = "Plamen Zahariev and Yahya Tashtoush and Omar Darwish",
booktitle = "2025 International Conference on Cybersecurity and AI-Based Systems, Cyber-AI 2025",
address = "United States",
}