PatentVector Information

PatentVector employs network analytics (eigenvector centrality and hierarchical graphing) and machine learning to build an influence-weighted comprehensive map of the worldwide patent system, currently encompassing more than 150 million patents and patent applications, with about 400 million interconnections, from more than 200 countries and patent authorities. Derived from this network, each patent has its own objective PatentVector (PV) Score, with 1.0 being the mean. These scores scale directly, so PV Scores of 2.0 and 0.5 are twice and half as important as average, respectively. From these PV Scores we also automatically derive dollar estimates of patent value for all patents and patent applications. The structure of our patent network allows users to locate progressively specific natural technology clusters, which are comprehensive, and are district from, and more accurate than, hierarchically-assigned CPC or IPC classifications, or groups of patents located using semantic searches, Boolean searches, or machine language approaches. We support multifactorial by patent number, title, abstract, inventor, assignee, patent office classification (USPTO, IPC, and CPC codes), country/jurisdiction, and date range. We have analyzed individual patents, portfolios, technologies, and entire national patent systems. PatentVector has been named a top 10 disruptive legaltech company by Stanford’s CodeX, our network building algorithm won the 2016 WSDM Cup, and Parabellum Capital announced (in December 2018) that “Over the past six months, we’ve performed an exhaustive analysis of patent data analytics offerings and concluded that PatentVector was the clear leader.” PatentVector is available by subscription, and may also provide custom data or reports.