Experimental Results of Using Cognitive Maps of Knowledge Diagnosis in Different Modes to Explain the Recommendations of Intelligent Tutoring Systems

Decisions made by Intelligent Tutoring Systems (ITS) in e-learning should not only be appropriate to the learning situation but also be understood by the learner. This increases the level of acceptance of the recommendations offered by the system to the learner and provides additional data that increases the learner’s trust in the ITS decisions. The mapping mechanism we used (Cognitive Maps of Knowledge Diagnosis, CMKD) allows us to organize the visual accompaniment of the dialog. In this research, we used different types of CMKDs and their display modes to interact with learners and evaluated their impact on the credibility (explainability) of the tutoring system’s decisions in the context of the cross-cutting approach to learning situation analysis. It was experimentally shown that particular and simplified maps are most effective under different strategies for conducting explanatory dialog with ITS. These conclusions are consistent with the results of the questionnaire.