Document Details

Document Type : Thesis 
Document Title :
Student’s Behavior Analysis to Recognize the Engagement Level
تحليل سلوك الطالب للتعرف على مستوى الانخراط
 
Subject : Faculty of Computing and Information Technology 
Document Language : Arabic 
Abstract : In the last few decades, adaptive e-learning systems have generated tremendous interest among researchers in computer-based education. Measuring student's engagement is an important key to improve adaptive e-learning systems. An e-learning system adapted to learner emotions was considered as an innovative system. Among the challenges that face researchers is how to measure student's engagement depending on their emotions. During a few years, several solutions were proposed to measure student's engagement, but few solutions are behaviors-based. Thus, this thesis aims to propose a new solution to increase the accuracy of measure student's engagement that relies on behaviors. According to our survey, all so-far proposed solutions for measuring student's engagement are based either on Self-reports and Observational checklists, monitoring student’s emotions or on physiological and neurological sensor readings. There has been an increasing interest in computer vision and camera-based solutions as a technology that overcomes the limits of human observations and expensive equipment involved for student’s engagement measurement. In this thesis, we propose and validate a new engagement affective model that links between engagement level and emotions. Furthermore, we propose an automatic multimodal approach to measure student's engagement in real-time based on computer vision tools. Thus, to provide more robust and accurate student’s engagement measurement, we combine and analyze three modalities representing student's behaviors: (1) emotions from face expressions, (2) keyboard keystroke, (3) and mouse movement. Such a solution operates in real-time while offering the exact level of engagement and using the least expensive equipment possible. We validate the proposed multimodal approach through three main experiments: single, dual, and multimodal on new Engagement-Datasets. We built new and realistic student engagement-Datasets to validate our contributions. We record the highest accuracy (95.23%) with a multimodal approach and the smallest MSE “0.04” compared to single and dual modalities. 
Supervisor : Dr. Salma Mohamad Kammoun 
Thesis Type : Master Thesis 
Publishing Year : 1441 AH
2020 AD 
Co-Supervisor : Dr. Arwa Abdulaziz Allinjawi 
Added Date : Wednesday, June 3, 2020 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
خولة عبد الرحمن الطويرقيAltowairqi, Khawlah AbdulrahmanResearcherMaster 

Files

File NameTypeDescription
 46280.pdf pdf 

Back To Researches Page