Abstract:
Social networks reveal the web of connections that exist among individuals.
The meaning of these relationships, however, remains largely unspecified. Generally, automated techniques fall short when it comes to identifying their nature. This paper discusses the use of crowdsourcing techniques to leverage expert knowledge for classifying social relationships derived from digital archival materials. In order to enhance the semantics of social relationships with contextual information (e.g., space and time), an application is proposed that relies on human annotation driven by dynamically-generated
questions. The system has the potential to provide scholars with a tool to support their analysis and enhance their understanding of the structure and dynamics of social groups and communities.
