InnovationQ Tips: Using Semantic to Enhance Boolean Searching

InnovationQ Tips Tags: Boolean Search, Concept Search, InnovationQ Tips, Keyword Search, Semantic Search

By: Nasreen Bakht Brady, Client Engagement Director, IP.com

I’ve written in the past about the mindset shift necessary to go from traditional Boolean/keyword searching to semantic/concept searching.  AI-driven technology in search is here, but embracing semantic doesn’t necessarily mean abandoning Boolean.  You can use InnovationQ’s semantic capabilities to enhance your Boolean searching.

Semantic search can be coupled with Boolean in a variety of ways to gain efficiency and thoroughness.  Here are three of the most accessible ways to use semantic to improve Boolean searching, even if you haven’t quite embraced semantic search.

1. Use a semantic search to help quickly learn about a new technology and find additional keywords to use in Boolean queries.

Method: Run a semantic search with a simple natural language description of what you’re looking for – even a broad concept, – or cut and paste information readily available about the area of interest (e.g. invention disclosures) as the main concept.  You may also use an on-point document as the main concept.  View top results, visualizations including Concept Term Heatmap/Cloud and Semantic Map to understand the space and learn important keywords, which can be used in Boolean queries.  In InnovationQ, the concepts prevalent in the most relevant results can be viewed in visuals or highlighting.

This is especially helpful in new or unfamiliar areas with unfamiliar terminology.  Broad concept searches can help teach about an unfamiliar space and uncover different segments which you may wish to focus follow-up searches.  This method can be a fast and easy first step to aid in crafting queries.

2. Use a semantic search to quickly identify relevant CPCs/IPCs and/or Assignees/Inventors to be used for your Boolean query and a batch-reviewing.

Method: As in number 1, above, enter the main concept for a semantic search using natural language, cut and pasted text, or a document. Switch to visualizations, and review appropriate charts.

Charts are shown by relevance to teach not only who the biggest players or most common CPCs/IPCs are, but also which are linked to the most relevant documents.  This is important to avoid missing small but important players or classifications, which can be used as a Boolean filter.

The charts are dynamic for visual searching.  For example, in the CPC by relevance chart shown, click on individual CPCs and add them as a filter.  Conveniently filter to certain CPCs from the chart and consider adjusting the query.  These filters may also be helpful in reviewing results in batches.

3. Use semantic search with a Boolean search to review results in order of relevance.

Method: Enter a Boolean query as usual and add the most important keywords as the main concept. Adding even a broad concept will apply relevance scores to the results, as well as visuals which consider relevance or results.

A difficulty in Boolean searching is that you must review all the results, whether they are first or last on the result set.  This means that the number of results returned is extremely important, and can necessitate adding keywords and other filters to reduce the results to a manageable number.  It also means you ought to approach reviewing results with the same gusto from start to finish, and that can be time-consuming.  By adding a concept, the results will be sorted by relevance so that you may review results, starting with the most important.  Not only can you focus your time on the most relevant documents, you may find a point down the list where the results are no longer (as) relevant and reduce the review time below that point.

Stay tuned to learn more about the power of combining semantic and Boolean search. If you’re interested in our upcoming webinar on November 15, “The Power of  Two: Combining Concept and Boolean Search,” just click on one of the links below to register.