A Dialogue Annotation Scheme for Weight Management Chat using the Trans-Theoretical Model of Health Behavior Change
This work addresses the problem of automating health behavior analysis in dialogues for weight management, but it is incremental as it focuses on annotation and classification without major methodological breakthroughs.
The study collected and annotated human-human role-play dialogues for weight management using a novel scheme based on the trans-theoretical model, and built classifiers to predict labels, finding that accuracy improved with oracle sentence segmentations compared to unsegmented sentences.
In this study we collect and annotate human-human role-play dialogues in the domain of weight management. There are two roles in the conversation: the "seeker" who is looking for ways to lose weight and the "helper" who provides suggestions to help the "seeker" in their weight loss journey. The chat dialogues collected are then annotated with a novel annotation scheme inspired by a popular health behavior change theory called "trans-theoretical model of health behavior change". We also build classifiers to automatically predict the annotation labels used in our corpus. We find that classification accuracy improves when oracle segmentations of the interlocutors' sentences are provided compared to directly classifying unsegmented sentences.