A Proxy-Based Method for Mapping Discrete Emotions onto VAD model (2511.12521v1)

Abstract: Mapping discrete and dimensional models of emotion remains a persistent challenge in affective science and computing. This incompatibility hinders the combination of valuable data sets, creating a significant bottleneck for training robust machine learning models. To bridge this gap, this paper presents a novel, human-centric, proxy-based approach that transcends purely computational or direct mapping techniques. Implemented through a web-based survey, the method utilizes simple, user-generated geometric animations as intermediary artifacts to establish a correspondence between discrete emotion labels and the continuous valence-arousal-dominance (VAD) space. The approach involves a two-phase process: first, each participant creates an animation to represent a given emotion label (encoding); then, they immediately assess their own creation on the three VAD dimensions. The method was empirically validated and refined through two iterative user studies. The results confirmed the method's robustness. Combining the data from both studies generated a final, comprehensive mapping between discrete and dimensional models.