The exploration of knowledge networks through patent citations

The aim of this chapter is to discuss the methodological issues arising from the use of patent citations data as relational data and to provide an example of empirical analysis. In particular, we propose to use patent citations data to explore knowledge networks at the level of organizations, that is, companies and other institutions. We focus our analysis on two different types of network: citation and co-citation networks. We argue that the former type of network could map knowledge diffusion networks; the latter could map the technological space. Using this perspective, this chapter examines the main structural properties of the knowledge networks in semiconductor and computer industries. In order to capture the organization's position in both co-citation and direct patent citation networks, we calculate, for each sector, two different measures: the crowdedness of a company's position in the technological space and a measure of its status. These two measures allow us to map the position of any organization and to defi ne a taxonomy of four different kinds of organizations: technological leaders, technological followers, brokers of new technologies and isolated organizations. The results of our analysis shows differences between the semiconductor and computer sectors and the difference between European and non-European companies in terms of technological niches.