Visualization and Critical Evaluation of Automatically Generated Concept Maps


  • Nika Marija Rojc University of Ljubljana
  • Andrej Košir University of Ljubljana


One of the persisting challenges in education is the task of effectively conveying knowledge to students and learning the materials as a student. Appropriate study techniques are essential for students to be in charge of their learning process, and it has been found that using multiple study techniques can significantly improve academic achievements [1]. 

A concept map is a schematic network of concepts that shows information organized hierarchically and it includes the relationships among different components of a given study topic in a graphic form. Concept mapping is one of the several strategies a higher education student can use to accomplish their study objectives. It involves visually organizing and connecting complex concepts which can improve comprehension of the subject [2]. By using a combination of natural language processing tools, linguistic and statistical algorithms, machine learning models, and graph theory principles, we can create systems that can be used to automatically generate concept maps [2]. 

The main purpose of this project is to experimentally create an automatically generated concept map using natural language processing libraries. There is no optimal tool for keyword extraction, therefore several methods will be applied. The chosen study material from the field of cognitive psychology will be preprocessed and made into a concept map. We will extract the main concepts as well as the relationships between them and then graphically visualize the results. A separate, manually generated concept map will be made and used to assess the quality of the automatically generated one. Both developed concept maps will be evaluated by experts using a set list of criteria that consider the quality of concepts and ideas, relationships between concepts, hierarchy, propositions, and spelling [1].

We aim to gain insight into the potential benefits of the automatic generation of concept maps as an additional study resource for higher education students.


[1] C. Romero, M. Cazorla, and O. Buzón, “Meaningful learning using concept maps as a learning strategy,” Journal of Technology and Science Education, vol. 7, no. 3, p. 313, Sep. 2017, doi.

[2] V. D. Santos et al., “Conceptual Map Creation from Natural Language Processing: a Systematic Mapping Study,” Revista Brasileira de Informática na Educação, vol. 27, no. 03, pp. 150–176, Dec. 2019. doi: