Abstract
When studying patent data as a way to understand innovation and technological change, the conventional indicators might fall short, and categorizing technologies based on the existing classification systems used by patent authorities could cause inaccuracy and misclassification, as shown in literature. Gao et al. (International Workshop on Complex Networks and their Applications, 2017) have established a method to analyze patent classes of similar technologies as network communities. In this paper, we adopt the stabilized Louvain method for network community detection to improve consistency and stability. Incorporating the overlapping community mapping algorithm, we also develop a new method to identify the central nodes based on the temporal evolution of the network structure and track the changes of communities over time. A case study of Germany’s patent data is used to demonstrate and verify the application of the method and the results. Compared to the non-network metrics and conventional network measures, we offer a heuristic approach with a dynamic view and more stable results.
Highlights
Patent data has attracted the interest of researchers as a way to measure and understand innovation and technological change, especially with the increased availability of online electronic database and the efforts made by worldwide patent authorities to consolidate and harmonize patent data at international level (Maraut et al 2008; OECD 2009).Gao et al (2017) have introduced an approach to construct networks based on the OECD Triadic Patent Family database (Dernis and Khan 2004), to identify communities and the community cores
Incorporating the overlapping community mapping algorithm, we develop a new method to identify the central nodes based on the temporal evolution of the network structure and track the changes of communities over time
Pharmaceutical products in International Patent Classification (IPC) class A61 and enzymology or microbiology in class C12 frequently co-occur in patent families and they are found to be in the same network community
Summary
Patent data has attracted the interest of researchers as a way to measure and understand innovation and technological change, especially with the increased availability of online electronic database and the efforts made by worldwide patent authorities to consolidate and harmonize patent data at international level (Maraut et al 2008; OECD 2009).Gao et al (2017) have introduced an approach to construct networks based on the OECD Triadic Patent Family database (Dernis and Khan 2004), to identify communities and the community cores. Most of the well-recognized conventional indicators are straightforward measures, such as the number of patent applications and publications, time needed from filing to grant (grant lag), number of different technology classification codes involved (patent scope), forward and backward citation counts, etc. Such indicators may be used to track technological changes and innovation, but when considered alone, will fall short due to their simplicity and lack of context, resulting in bias and sometimes contradicting conclusions (Benner and Waldfogel 2008; Dang and Motohashi 2015; Hall et al 2001; Hall and et al 2005; Harhoff et al 2003). Pharmaceutical products in IPC class A61 and enzymology or microbiology in class C12 frequently co-occur in patent families and they are found to be in the same network community
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
