Linguistic and Vocal Markers of Microbehaviors Between Team Members During Analog Space Exploration Missions

Abstract

We used machine learning classifiers and dialog state tracking models, combined with natural language processing techniques relying on lexicon-based methods and data-driven methods, to automatically detect positive and negative microbehaviors between team members in nine, four-person teams in a simulated space habitat. Our findings indicate that the psycholinguistic markers extracted using the linguistic inquiry and word count, STRESSnet dictionaries, and acoustic features can achieve an f1-score up to 54.87% in a three-class classification problem. Our findings also suggest that modeling turns between the sender and target of microbehaviors is significantly more effective in detecting microbehavior than only modeling the sender’s information. Finally, we demonstrate the effect of introducing context for detection purposes. Dialog state tracking approaches that model the linguistic interaction between team members and incorporate contextual information about the task and sentiment of the conversation can further yield improved performance, depicting an f1-score of 57.73%.