Vector Space

Patent Analysis

- in Real Time

Companies who use our vector space analytics obtain the highest level of confidence about their patents and patent strategy, in near real time, reducing risks and costs.

The Patent Research Foundation is the only company to provide an extensive vector space analysis of each and every patent, because doing so requires the combination of very disparate and specialized areas of expertise:

Scientific computing

The scholars at the Patent Research Foundation are at the forefront of scientific computing and big data technologies. Our analytics tool calculates a high-dimensional Vector Space Model with over 700,000 dimensions to cover the entire known technology space. We then start from a point of reference that you provide (such as an existing patent, a patent application, or the text of a white paper or article) and compare that to every other patent publication and patent grant at the USPTO, on every single dimension. Most importantly, we cross-check every single word in the space to identify associations that are hard to find. Because we publish our methodology, our analytics are open, scientifically validated, and peer reviewed, giving you results that you can trust. With other companies, there is little or no transparency - you get buzzwords and “black box” descriptions that have not been validated and are difficult to trust.

Patent prosecution and patent data

Even the most advanced technology, when it is applied in a generic way, will produce poor or misleading results. Reliable analytics requires deep expertise about the data being analyzed - and the institutional processes that generate that data. Patent attorneys and scholars at PRF have accumulated years of experience working with the USPTO, and have co-authored papers with other top academics in the field. Most patent analysis companies blindly rely on data from second-hand providers, further introducing error into their analysis. Our data go through numerous corrections, based on steps validated by domain experts, to get the analysis right.

Big data processing

Data engineering for scalability and speed of execution is a very different kind of problem from figuring out the right model to run in the first place. Competitors rely solely on topic modeling methods such as Semantic Gist Analysis, Semantic Search, Latent Derelict Analysis, Latent Semantic Analysis, Semantic Boolean Analysis, and others, which collapse the search space down to make computation much, much easier. Although topic modeling tools are easy to apply off-the-shelf, the results can be unpredictable and unreliable. In contrast, we have built an engine that compares every single word, in every issued patent or published patent application, to every other word in every other patent document since 1976. We do that across 700,000 dimensions in near real time. Leveraging scalable cloud computing resources, we process over 295 terabytes of data on a distributed computer that scales up to the size of a city block.

Strategic positioning

When you work with PRF, we’ll help you use our reports to complement your existing processes to make them more efficient, more reliable, and more supportive of an overall patent strategy. Scholars at the Patent Research Foundation also research and teach at leading business schools - that gives us a unique view on how to use our patent analytics to support your firm’s strategy. We meet - virtually or physically - with every new customer account to make sure PRF works for you.

Having expertise in only one, or perhaps 2, or even 3 of the above areas is not sufficient. Only PRF has the deep level of expertise across all four disciplines to give you results, in near real time, that you can trust.

Contact us to learn how your company can achieve the superior results from our analytics.