Measuring the Similarity of Concept Maps According to Pedagogical Criteria

A concept map provides a graphic hierarchical means of representing how knowledge is structured in a domain. It visually organizes a set of concepts showing their mutual relations. An analysis of the similarities between two concept maps can produce significant results in any fields where Intelligent Knowledge Management is used, such as Healthcare, Policy Development, Energy and Waste Management, Resource Consumption Sustainability, Mobility, Safety, Citizen Empowerment, and, of course, Education. In an educational setting, a concept map conveys the various concepts connected by relations of dependence that a course must cover. However, the similarity between two concept maps for education has to be measured according to criteria that take into consideration the pedagogical properties of the maps, i.e., not only considering the structural aspects of the maps themselves. An automated analysis of the similarity between two concept maps can allow the teacher to reflect on different interpretations of the knowledge domain of a certain course as well as to assess how existing learning material can be implemented in a new course. Research into this aspect of concept mapping appears to be relatively scarce. This paper proposes criteria to assess the similarity of two concept maps, also based on pedagogical features, with the aim of providing teachers with better support during the course creation process. Each criterion is implemented through a specific measure function. The measures are then shown to be sensitive to their criterion rationale by evaluating them against a collection of random case studies.