Social-Behavioral Modeling for Complex Systems

21 Evaluation and Validation Approaches for Simulation of Social Behavior: Challenges and Opportunities

Emily Saldanha¹, Leslie M. Blaha², Arun V. Sathanur³, Nathan Hodas¹, Svitlana Volkova¹ and Mark Greaves¹

¹ Data Sciences and Analytics Group, National Security Directorate, Pacific Northwest National Laboratory, Richland, WA, 99354, USA

² Visual Analytics, Pacific Northwest National Laboratory, Richland, WA, 99354, USA

³ Physical and Computational Sciences Directorate, Pacific Northwest National Laboratory, Seattle, WA 98109, USA

Overview

Broad Observations

Humans often behave in ways that defy simple explanations. For this reason, complex behavior models may be used to draw together what we know and believe about humans and to help us model and predict events in the real world. For example, we may like to know the effect of a public health intervention on diabetes rates (Jones et al. 2006), forecast human migration during a refugee crisis (Edwards 2008), or understand how online content becomes viral (Hodas and Lerman 2014). As we advance the science of modeling and simulation of complex social systems, we are challenged to grapple with complexity at multiple levels of the system. Individual decision‐making behaviors can be modeled, with complexity of behavior emerging from basic cognitive mechanisms. The interpersonal interactions between individuals then give rise to the complex, dynamic behaviors of social systems. It follows that the development of new ...