Detecting innovation in networks of collaboration: A graph theoretical approach

Innovation in science is difficult to define other than in hindsight. In this paper I present a novel method to identify and analyze patterns of scientific innovation based on advanced work in graph theory. Widely accepted innovations lead to a subsequent rearrangement of patterns of collaboration that is detectable by signal detection algorithms applied to time series of large networks of collaborations. This method enables us to find hotspots of scientific activity amendable for further analysis. To pursue this kind of interdisciplinary research effectively also requires us to develop a novel type of research system for data driven computational approaches at the intersection of science, history of science and science policy.


Part of session

Big Data - Smart Mining

Related documents