Investigating Clues for Estimating ICAP States Based on Learners' Behavioural Data During Collaborative Learning

摘要

Interactions based on the learners' state of understanding and their attitudes toward tasks are considered important for realising a support system for collaborative learning. In this study, as a first step, we tried to detect whether the learner's state is Passive in the ICAP theory from the data obtained during collaborative learning. We actually conducted an experiment of collaborative learning between participants and obtained data on facial features, gaze directions, and speech state during the experiment. Based on these data, we investigated clues to classify the status of ICAP as either Passive or not. As a result, we were able to find several candidates. On the other hand, in the state classification of participants' states using these independent variables, it was not possible to show high accuracy. In future experiments, we plan to simultaneously measure physiological indices as a clue to estimate participants' internal state.