Analysis of concentration level of participants toward development of meeting assist system

Abstract

Meetings are important for various decision-making processes in organizations. On the other hand, longer meetings are stressful for the participants and cause a decline in their concentration level. The purpose of this research is to construct a model that analyzes the concentration state of meeting participants. In this research, we obtained meeting data for a total of 17 hours and 30 participants with a variety of agenda settings. From the recorded data, we annotated speech, nods, and smiles. We then added four scales of atmospheric features estimated from the speech information to the seven explanatory variables, and constructed a logistic regression model to analyze the concentration state of them, which was defined based on the TotalHb value. By applying the results of this research, it is expected to be possible to provide feedback according to the estimated concentration state of the meeting participants, thereby realizing more productive meetings.