|
Patent information is essential - but getting it can
prove to be deeply problematic
To improve transparency, patent offices of the leading industrialized countries
provide digital patent data. This is the good news. The bad news is, that the
data is not structured nor standardized. Furthermore, patent data of national
and regional patent offices are not linked. Patent offices and patent attorneys
analyze patent data bases and open-source technology before granting new patent
applications. The industry analyses patent data to monitor competitor's patent
activity and industry trends, to decide on R&D investments, to evaluate
potential infringements etc. To find patent data, corporations and patent
attorneys use software tools or hire costly services.
THE LIMIT OF BOOLEAN KEYWORD SEARCH
Various players offer patent search software tools and services. All tools and
services currently available are based on keyword (Boolean) search technology.
However, keyword search results often miss relevant patents. Even the best
patent analysis can be dangerous when based on insufficient patent information.
Misleading assumptions can result in wrong decisions, billions of mis-spend R&D
investment and unrealized revenue opportunities.
Keyword search technology is simple and returns only documents containing the
selected keyword. It requires an exact match on search terms. The results depend
on the quality of keyword selection and the quality of the database. Keyword
allows for a summarized index to the documents by category, but there's a
drawback: the sheer number of patent documents to be summarized; and, if a
document is not properly summarized, it can be returned as a false positive - or
skipped, despite being relevant. It is good for true/false searching or very
clear data searching (i.e., does the word exist). But to hide R&D strategies,
many patents are purposely written to obscure discovery (new lexicon). They do
not contain obvious keywords and will be missed. With Boolean search, a small
number of keyĂords return a high number of false-positive hits, which demands
labour-intensive manual analysis. High numbers of keywords skip high numbers of
patents, even if they're relevant. Scanning every single word of all documents,
keyword efficiency is limited when searching relevant data of over 27 million
patents (1/4 billion pages of full text).
|
