A latent space model for cognitive social structures data

This paper introduces a novel approach for modeling set of directed, binary networks in the context cognitive social structures (CSSs) data. We adopt relativist which no assumption is made about existence an underlying true network. More specifically, we rely on generalized linear model that incorporates bilinear structure to transitivity effects within networks, and hierarchical specification borrow information across networks. spatial model, perception each individual strength relationships can be explained by perceived position actors (themselves others) latent space. A key goal provide mechanism formally assess agreement between actors’ their own roles with rest group. Our experiments both real simulated data show capabilities our are comparable or, even superior to, other models CSS reported literature.