ABSTRACT
In this chapter we focus on multimodal analytics, which is an emerging field concerned with designing automated assessments of constructs which are difficult to assess using traditional methods (e.g., collaboration, interview skills, emotion regulation, attentional lapses). At its core, multimodal analytics involves the integration of multiple modalities (e.g., audio and video) via machine learning or some other statistical method to automate the assessment of complex constructs in varied settings. We provide a broad overview of this emerging field by reviewing basic principles, computational methods, a variety of case studies, and guidelines to consider in designing such systems.
