Methods for automatically analyzing conversational turn-taking patterns

摘要

Presented are a system and methods for automatically extracting, analyzing and visualizing conversational turn-taking patterns during co-located and remote social interactions. The visualization tool measures group conversation dynamics based on wirelessly or otherwise obtained electronic sensor data corresponding to each participant. The group conversation dynamics that are determined (or detected) include, per participant, turn duration, turn-taking speed, number of turns, number of overlapping turns, number of successful interruptions and number of unsuccessful interruptions. Also calculated are participant mirroring statistics, activity statistics, consistency statistics, influence statistics and social network statistics. The visualizations may be output with user configurable parameters.