Using Simple Speech–Based Features to Detect the State of a Meeting and the Roles of the Meeting Participants

We introduce a simple taxonomy of meeting states and participant roles. Our goal is to automatically detect the state of a meeting and the role of each meeting participant and to do so concurrent with a meeting. We trained a decision tree classifier that learns to detect these states and roles from simple speech–based features that are easy to compute automatically. This classifier detects meeting states 18% absolute more accurately than a random classifier, and detects participant roles 10% absolute more accurately than a majority classifier. The results imply that simple, easy to compute features can be used for this purpose.