Optimisation des systèmes de gestion de l'apprentissage par l'intelligence artificielle

Thesis Abstract

The rapid proliferation of digital learning environments has underscored the necessity for intelligent and adaptive learning management systems (LMS) that can effectively meet diverse educational needs. Despite significant advancements, current LMS often lack personalization capabilities and adaptive features critical for enhancing learner engagement, retention, and academic performance. This study aims to optimize LMS functionalities through the strategic integration of artificial intelligence (AI) techniques, thereby advancing the effectiveness and efficiency of digital education platforms. The primary objective is to develop an AI-driven framework that personalizes learning experiences, predicts learner performance, and automates administrative tasks within LMS environments. To achieve this, the research adopts a mixed-methods approach, combining quantitative and qualitative data collection strategies. The quantitative component involves a survey administered to 300 university students enrolled in online courses across three institutions, alongside an analysis of system logs from 10 different LMS implementations. Data collection instruments include a structured questionnaire measuring learner satisfaction, engagement, perceived system effectiveness, and demographic variables, complemented by system-generated data capturing user interactions. Qualitative data will be obtained through semi-structured interviews with 20 instructors and system administrators to explore contextual factors influencing AI integration and system optimization. The analysis employs multiple regression analysis to identify key predictors of learner success, factor analysis to validate constructs related to system usability and perceived effectiveness, and thematic analysis for qualitative interview data. Additionally, machine learning techniques such as supervised classification algorithms (e.g., Random Forest, Support Vector Machine) will be used to predict learner outcomes based on interaction patterns, while optimization algorithms like genetic algorithms will be employed to enhance system parameter tuning. Expected findings suggest that AI-driven personalization features significantly improve student engagement and academic achievement by tailoring content delivery and assessments based on individual learner profiles. The predictive models are anticipated to accurately identify learners at risk of dropping out or underperforming, allowing targeted interventions. Moreover, automation of administrative functions through AI is expected to reduce system management costs and increase operational efficiency. The integration of theoretical models such as Vygotsky’s Social Development Theory and the Cognitive Load Theory will underpin the framework, illustrating how AI facilitates scaffolding learning experiences adapted to learner-specific cognitive loads and social contexts. This research intends to contribute substantially to the emerging body of knowledge on AI-enhanced educational technologies by proposing a comprehensive, empirically validated framework for LMS optimization. It underscores how intelligent systems can transcend static content delivery and evolve into proactive, adaptive learning environments. The findings will provide practical insights and guidelines for policymakers, educational technologists, and system developers aiming to implement AI solutions effectively within diverse educational settings. The study concludes that embedding AI capabilities into LMS significantly enhances teaching and learning dynamics, fostering more personalized, engaging, and effective learning experiences. Recommendations include establishing institutional policies for AI integration, investing in training for educators on AI functionalities, and advancing research on ethically aligned AI deployment in education. Future studies are suggested to explore longitudinal impacts of AI-driven LMS on academic performance and to investigate the potential of emerging AI paradigms—such as deep learning and natural language processing—in further refining adaptive learning environments.