Keyword Weighting in FreePatentsOnline and Delphion

[tweetmeme source=”Intellogist” only_single=false]

Keyword weighting is a search technique that can help the user hone in on specific concepts during an otherwise ordinary keyword search. Patent search systems where keyword weighting is available include FreePatentsOnline, SumoBrain, and Delphion. As seen in last week’s TotalPatent Semantic Search post, keyword weighting can also be used in patent analysis tools to narrow the focus on certain areas of subject matter. Often, certain keywords are more important to the searcher than others. One keyword may be more central to the concept and another may be linked to an elusive subject feature. By weighting specific keywords, searchers can bring the best results to the top of the pile.

Keyword search using weighted terms is a method available only available in search systems that also employ a ranked results list. The results list may also be known as “weighted” or sorted by “relevancy” depending on the search engine. Weighting keyword terms allows users to specify that they should have different relative importance when the mathematical formula for relevancy is computed. The most common and basic formula for ranking a results list relies on keyword frequency. Keyword frequency merely counts up the number of keywords in a document (sometimes limiting the search to certain areas of a document), and in most cases, calculates a form of keyword density—the ratio of keywords to non-keywords. Other formulas for ranking a results list may have to do with keyword proximity, latent semantic analysis, or a proprietary ranking algorithm (Google, for instance).

FreePatentsOnline and SumoBrain employ keyword weighting as a supplement to their relevancy ranking method of listing search results. For more info on the similarities and differences between FreePatentsOnline and SumoBrain, see last week’s Intellogist Blog post on the subject. Users can use the “^” operator to specify weighting within any normal keyword search on FreePatentsOnline or SumoBrain. For example, a query of “wrench^4 OR nut” would return results of the query “wrench OR nut,” where “wrench” is 4 times more important to the relevancy score than “nut.” Similarly, keyword weighting can be accomplished in Delphion through use of numbered operators, as seen in this example: ([[100]](toothbrush and holder) or [[50]]holder). This search string will return documents with the keywords toothbrush and holder, and will also return documents containing only the term holder. Those documents containing both keywords will receive a higher relevancy ranking than just those containing holder.

A skilled searcher with experience using keyword weighting features can more quickly obtain relevant results than a searcher limited to using brute force keyword searches. Using weighting and ranking systems can shorten the time spent searching for a particular feature, since the highly ranked results appear at the top and are more likely to reward searchers with prior art. The rise of Google has proven that users want relevant information quickly. My experience with patent searching and the patent searching community have taught me that users want fine control over how search strings are crafted and entered. Keyword weighting is one solution to bridging both of these gaps.

Do you have any experience with keyword weighting (in a patent searching context or not)? Let us know what you think in the comments below!

Patent Searches from Landon IP

This post was contributed by Intellogist team member Chris Jagalla.


One Response

  1. […] include AND, OR, and NOT. More advanced operators in any given system may include proximity and keyword weighting varieties (click on the links to see earlier posts on these subjects). Today I’ll highlight […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: