Identifying spatial technology clusters from patenting concentrations using heat map kernel density estimation

Abstract

In this paper a methodology for identifying and delineating spatial technology clusters based on patenting concentration is developed. The methodology involves the automated geocoding of patent inventor addresses, the application of a home bias correction factor and a sensitivity analysis to determine the optimal parameters of the kernel density estimation interpolation distance and the minimum concentration threshold to identify clusters. The methodology’s performance is compared to a number of other cluster identification methods and it is validated across 18 individual sectors, including mature broad-based high-technology sectors and emerging niche sustainable energy technology sectors. The results suggest that the performance of the methodology exceed that of alternative cluster identification methods, although there is some variation in performance between different sectors. This demonstrates that the methodology provides researchers, practitioners and policy makers with a useful tool to gain insight into the spatial distribution of sectoral innovation activity at a global scale and sub-national regional level and to monitor changes over time, thereby supplementing more readily available global statistical data which is available at the national level.