What the Fuzz!? Leveraging Ambiguity in Dynamic Network Models

In a perfect world, data could be placed neatly into discrete categories; however, the world we live in is far from ideal. A consistent finding from studying atomic bonds to psychopathy is that variables and people seldom cooperate with our attempts to fit them into neat clusters. The argument for the current work is that ambiguities in group membership may be more informative than we acknowledge and warrant adoption of “fuzzy” methodologies that can explicitly quantify these uncertainties. Here, we describe fuzzy community detection methods which permit variables to belong to multiple communities by varying degrees rather than restricting their membership to “on” or “off” and contrast them with discrete forms of community detection. In line with recent calls for dimensional approaches to the study of psychopathology, we discuss how the application of these fuzzy methodologies may inform the study of psychological symptoms and subjects. In particular, we discuss how ambiguous symptoms could serve as the bridges between psychopathologies and how ambiguous subjects could exhibit dynamics relevant for treatment.