Development of an Intelligent Knowledge Mangement System for Engineering Education

Abstract

The growing complexity of engineering education and the rapid expansion of digital resources have created new challenges in managing, retrieving, and personalizing access to knowledge. This thesis presents the design, implementation, and evaluation of an intelligent knowledge management system developed to address the specific needs of engineering education. The proposed system integrates semantic search, graph-based recommendations, and modular learning paths to support structured, user-centered learning experiences. The platform architecture combines a PostgreSQL database, Elasticsearch for semantic retrieval, and Neo4j for personalized recommendations based on user interactions and content relationships. A mixed-methods evaluation was conducted to assess technical performance, usability, and educational relevance. Results indicated that the semantic search feature enabled more context-aware information retrieval beyond basic keyword matching (MAP = 0.627, Recall@5 = 0.714). The recommendation system provided relevant suggestions across multiple contexts within the platform, with consistently high precision in aligning recommendations to user interests and learning goals. Qualitative feedback also highlighted the usability of the platform, the clarity of the interface and the perceived educational value. However, the evaluation was subject to certain limitations, including the scarcity and limited diversity of available learning materials within the system, as well as the lack of a long-term assessment to measure learning outcomes over time. Despite these constraints, the findings illustrate the potential of intelligent technologies to support inclusive, adaptive and learner-centered educational environments aligned with the principles of Education 5.0.