Affect interpretation from multithreaded online conversations is a challenging task. Understanding context and identifying target audiences are very crucial for the appropriate interpretation of emotions implied in an individual input embedded in such online social interactions. In this paper, we discuss how context is sued to interpret affect implied in conversational inputs with weak affect indicators embedded in multithreaded social interactions. Topic theme detection using latent semantic analysis is applied to such inputs to identify their discussion themes and potential target audiences. Relationships between characters are also taken into account for affect analysis. Such semantic interpretation of the dialogue context also shows great potential in the recognition of metaphorical phenomena and the decelopment of a personalized intelligent tutor for drama improvisation.
|Title of host publication
|Intelligent Tutoring Systems
|Published - 2012
|Lecture Notes in Computer Science