Didik Dwi Prasetya, Ricky Tirta Hardiyanta, Azlan Mohd Zain, Tsukasa Hirashima
Advancements in AI and machine learning have introduced innovations in education, including the use of concept maps. Concept maps help identify gaps in understanding between learners and educators by analyzing text similarity. However, traditional methods for measuring text similarity are often inefficient. This study employs FastText, an NLP-based machine learning model, to analyze text similarity patterns in a corpus of concept maps from a relational database course. The FastText model was trained on the collected corpus to generate deep vector representations of words, enabling more accurate similarity measurements. Results demonstrate that FastText outperforms traditional methods, with potential applications in personalized learning recommendation systems, cross-course concept mapping, and analysis of conceptual understanding development. The best performance was achieved at a threshold of 0.9822 with 93% accuracy and balanced evaluation metrics, demonstrating the effectiveness of FastText in identifying proposition similarity in open-ended concept maps. This research highlights the importance of integrating AI to enhance the effectiveness of educational tools like concept maps. © 2025 IEEE.
Universitas Negeri Malang, Department of Electrical Engineering and Informatics, Malang, Indonesia; Universiti Teknologi Malaysia, Faculty of Computing, Johor Bahru, Malaysia; Hiroshima University, Graduate School of Advanced Science and Engineering, Hiroshima, Japan
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock