Effectiveness of Automated Grading Tool Utilizing Similarity for Conceptual Modeling.

In recent years, a system engineer is required for advanced modeling because systems have become more complicated. A lecture of model diagrams has been carried out in many universities for training system engineers. The lecture has the problem that teachers cannot give enough feedback to students. The problem is due to a high degree of freedom of the model diagram. The model diagrams have words that have equal meaning and have different expressions. Teachers should consider these features when grading deliverables of students. In addition, a common way to grade deliverables is that teachers check "model diagrams" by hands. So, teachers need a lot of time when grading. Students cannot review their own deliverables. In this research, we develop a tool that gives feedback for students in lectures. We focus on conceptual modeling using notation based on class diagrams. Also, the tool automatically giving feedback for students at near real time. Teachers judge words that have equal meaning and have different expressions. The tool uses these judgements when grading deliverables of students. The content of feedback is a result of grading all operations. The results are visualized by a line chart. We experimented and verified the adequacy of the proposed method. We confirmed the usefulness of the method and the correlation between result of grading and time of unused KIfU.