Automatic Interpretation of Negotiators' Affect and Involvement Based on Their Non-verbal Behavior

Valid interpretation of the nonverbal behavior of the people involved in negotiations is important. Computational agents that are designed for negotiation benefit from the ability to interpret human nonverbal behavior for communicating more effectively and achieving their goals. In this paper, we demonstrate how the mode of involvement and relational affect of the negotiators involved in the interaction can be determined by several nonverbal behaviors such as that of the mouth, head, hand movements, posture and the facial expressions of the negotiators. We use machine learning to study involvement and affect in negotiation. Our results show that the prediction models built based on nonverbal cues can help identify the negotiator’s attitudes and motivation in the interaction.