Discussion of all things patent mapping and analytics.
English: United States Patent Cover from a real patent issued (Photo credit: Wikipedia)
Network Patent Analysis (NPA) and other citation based patent analysis methods are using lists of prior art documents, either prepared by patent examiners, or submitted by patent applicants.
One sometimes cited issue with this is that any patent examiner and even applicants are only human, like the rest of us, and can easily miss relevant patents due to inconsistent use of keywords or patent classification terms. But does mean that the results from citation analysis methods are flawed?
It depends on what citation analysis method you are using. In the case of NPA, not at all. The reason why is that rather than relying on any single examination report, NPA combines the search reports of every examiner that has examined a patent, and then distills down these combined search reports to identify the highest ranked (most connected) patents and the strongest connections.
Practically, this means that if a patent examiner or applicant does miss a relevant prior art document, you should still find it, as it is very unlikely that all patent examiners and applicants will miss it. And as long as at least one patent examiner or applicant can find the connection, you should still be in a position to find the patent.
So the answer to the question of how to be smarter than a patent examiner? Combine the search reports of many patent examiners - or in other words, make the most of their collective intelligence. And if this leads to a surplus of data, use techniques such as NPA to deal with this surplus of data.
The approach we take even considers indirect citation relationships. So in case that all examiners and applicants may miss a prior art invention, our algorithms may suggest relevant patents worth considering.