Predicting Persuasiveness of Participants in Multiparty Conversations

ABSTRACT Persuasiveness is an important capability in communication skills. This study aims to estimate the persuasiveness of participants in group discussions. First, human annotators rated the level of persuasiveness of each of four participants in group discussions. Next, multimodal and multiparty models were created to estimate the persuasiveness of each participant using speech, language, and visual (head pose) features using GRU-based neural network. The experimental results showed that multimodal and multiparty models performed better than unimodal and single-person models. The best performing multimodal multiparty model achieved 80% accuracy in predicting high/low persuasiveness, and 77% accuracy in predicting the most persuasive participant in the group.