Abstract

Understanding the topology of complex networks is a central concern of network science. Within this endeavor, we study the problems of building theories from the non topological attributes of linked vertices and assessing their explanatory power. We design a simple framework for building theories from the attributes of vertices and apply it to explain the topology of the Chilean shareholding network, an economic network which vertices represent firms and edges represent an ownership relation, finding that a relational theory based on financial information explained the topology of the network only in part.