Fine Tune Natural Language Queries

Natural language searching is a form of query by example or QBE, where you submit a block of source text, then ask the search engine to find documents similar to the example.  Rather that returning a short list of very targeted patents, natural language searches may return a million or more results, but the precision is in the sort or the ranking, where the first document is the best match, the second the second best, and so on down the list.

Natural language searching is commonly used to find similar patents, where the source text is a patent you may have found in AcclaimIP.  Another common usage is to paste in an invention description or a set of sample drafted claims to find documents that “read” like that block of text.

Many AcclaimIP users don’t use natural language searching correctly, and as a result they don’t use AcclaimIP to its full potential.  In this video, you’ll learn some simple techniques to refine your natural language searches, get better results, and find spot on prior art.

Natural language searching can be used in conjunction with keyword searching to help order your documents in a search result list.

While not covered in this video, when practicing with natural language searching, take a minute to view the filters and notice how the facets (elements in a filter) are derived from only the first 200 best matches.  This is a powerful technology that allows you to filter the set and still get matches you’d expect.  We call these SuperFacets.

In the video I reference another video called “How to do a Patent Keyword Search,” which you may want to watch to round out your keyword searching skills.

Post navigation