AI-LMS: AI-Based Behavioral Modeling of User's Interaction with the Learning Management Systems

  • Abdulrahman Awad

Student thesis: Master's Thesis

Abstract

Learning management system became a necessary part of the pedagogical phenomenon especially after the COVID-19 pandemic. Modelling of LMS users is vital to learners, instructors, policy makers and stakeholders. QoI is an effective metric that can model the user’s interaction with LMS which was proven across the different HEI levels. This work develops several deep learning models, mainly ANN to represent the expert knowledge encapsulated in the fuzzy logic system of Fuzzy-QoI allowing for generalizability and possibility of interpretation using SHAP. Deep models were interpreted, and force plots were used to provide feedback to users in order to improve their QoI. Additionally, two new datasets covering the quarantine period were collected from Moodle in Khalifa University, annotated, and prepared based on the work of Fuzzy-QoI. External validation using the new datasets provided insights into the advantages and disadvantages of the application of deep learning models in education. The feedback received from the models based on interpretation can improve both learners’ and instructors’ pedagogical experience while providing relevant insight to stakeholders and policymakers.
Date of AwardApr 2023
Original languageAmerican English
SupervisorLeontios Hadjileontiadis (Supervisor)

Keywords

  • Learning Management Systems
  • Quality of Interaction
  • Fuzzy Logic
  • High Education Institutes
  • Deep Learning
  • SHAP

Cite this

'