6ZE-04
Case Study on Evaluation of Students' Discussion Statements' Appropriateness Based on their Heart Rate
○彭 詩朦,大平茂輝,長尾 確(名大)
In this study, we explore how students’ Heart Rate (HR) data can be used to evaluate their answer-statements’ appropriateness while students completed a Question-and-Answer (Q&A) session in discussions. We adopt Apple Watch to collect students’ HR and Web-based scoring method to evaluate answer-statements’ appropriateness. HR features were analyzed and used in three machine learning models: logistic regression, support vector machine, and random forest for prediction of answer-statements’ appropriateness. Leave-one-student-out cross validation was used to evaluate classifiers’ accuracy on all of the students. We also take insight into the performance of HR-based prediction models on student groups with different level of experience on the discussion. We validated the effectiveness of our proposed models regarding evaluation of students’ discussion statements’ appropriateness.