The global structure of knowledge network

In this paper, we treat patent citations as knowledge networks connecting pieces of formalized knowledge and people, and focus on how ideas are connected, rather than how they are protected. We focus on the global structural properties of formalized knowledge network, and more specifically on the connective micro-mechanisms of network formation underlying such structures. Our main objective is to build the information infrastructure needed to address more focused questions about the global structure of innovation and knowledge production. Innovation and the creation of knowledge are increasingly understood as network phenomena, emerging from complex systems of dependence relations linking different social actors. The production of innovative ideas is best seen as collective enterprises involving transfer and recombination of existing knowledge through the interaction among people, resources, and institutions. Recent research recognizes the social character of knowledge production, where ideas are embedded in complex networks connected to other ideas. The large space spanned by knowledge networks is not homogeneous, and certain areas are more likely to produce innovation than others. Understanding how knowledge is produced by recombination of existing knowledge is the key to understand and predict technological, scientific and social innovation. This finds wide applicability in the analysis of patent data, which represent a clear example of how knowledge exchange becomes amenable for quantitative analysis. Against this general background, we construct what might be the largest and most complete knowledge network currently available, by merging separate patent datasets made available by the OECD, covering the last 40 years of global knowledge production. The resulted combined dataset reports patents data for 784,668 EPO, 690,227 PCT and 1,989,501 USPTO patents. Using this dataset, we extract the patent citation network that includes a total of 21,585,409 nodes and 111,233,700 ties, making it one of the largest longitudinal network datasets, and the largest knowledge exchange network. In this paper, we use graph-theoretic measures and measures developed in scientometrics to calculate the importance of patents, technologies, and inventors, and report the preliminary findings of our research project.