Patent data represent a significant source of information on innovation and the evolution of technology through networks of citations, co-invention and co-assignment of new patents. A major obstacle to extracting useful information from this data is the problem of name disambiguation: linking alternate spellings of individuals or institutions to a single identifier to uniquely determine the parties involved in the creation of a technology. In this paper, we describe a new algorithm that uses high-resolution geolocation to disambiguate both inventor and assignees on more than 3.6 million patents found in the European Patent Office (EPO), under the Patent Cooperation treaty (PCT), and in the US Patent and Trademark Office (USPTO). We show that our algorithm has both high precision and recall in comparison to a manual disambiguation of EPO assignee names in Boston and Paris, and show it performs well for a benchmark of USPTO inventor names that can be linked to a high-resolution address (but poorly for inventors that never provided a high quality address). The most significant benefit of this work is the high quality assignee disambiguation with worldwide coverage coupled with an inventor disambiguation that is competitive with other state of the art approaches. To our knowledge this is the broadest and most accurate simultaneous disambiguation and cross-linking of the inventor and assignee names for a significant fraction of patents in these three major patent collections.
Download