SHIELD (Novel Techniques for the Synthesis of High Fidelity Social Network Data)

This proposal is for developing Novel Techniques for the Synthesis of High Fidelity Social Network Data (SHIELD). The objective is to develop scalable methods, tools and techniques to synthesize and validate high fidelity social network data and to provide these to researchers in a modular, easy-to-use system. Our innovative program will integrate a set of social science tools for working with social networks; these include respondent driven sampling, consensus analysis, influence models, exponential random graph models, and the economy of relations. Using these tools, we will develop new scalable methods, tools and techniques to synthesize and validate high fidelity social networks which are similar enough to real networks such that applications developed with their data will work on corresponding real data. The essential features of these modeled networks are: they are longitudinal (i.e., they will account for change in nodes and links across time); they are dynamic (i.e., they will account for regular flows across relational interactions); their members respond to their dynamics; they involve multiple types of links, each with multiple attributes; and they involve multiple types of connected entities each with multiple attributes.