The Massive Differences between Keyword and Concept Searches
Keyword-based searches can help us quickly root out discrete facts, such as what time a store closes, who was the 34th U.S. president of the United States, or how to get to a local restaurant. But they are much less effective for finding more complicated details. Take, for instance, a simple search on patent claims.
Varying Results with Google Patent Keyword Search
The results returned on Google’s Patent site illustrate that how a query is worded impacts not just the number of results retrieved, but which results are included in each set. In the second example, the verb “lowers” is exchanged for the verb “reduces” while the third example swaps the term “aspirin” for its more familiar chemical name “acetylsalicylic acid” and “blood pressure” to its medical term “hypertension.” Each of these searches retrieves results that do not show up in the other searches, even though the claims are essentially the same except for the synonyms used.
Frighteningly, organizations have created entire workflows to accommodate the imprecision and redundancy in keyword-based search engines. Scientists begin by either searching every combination of the 202 synonyms of “aspirin,” 7 synonyms of “reduce,” and 19 synonyms of “blood pressure” (26,866 unique queries) or using a massive Boolean equation to build these queries. Often, each search must be run across separate databases with different protocols and syntaxes. And then each list of results must be reconciled to compile a master file of the 88,530 unique claims. So even though each result might come back in .1 second, typing in over 25 thousand searches and aggregating unique results can take weeks!
This is the type of research that scientists, researchers, stock analysts, national security officials and many other professionals have to do on a daily basis but on a much more complicated scale. This is why the issue of finding actionable insights from Big Data is in such high demand by so many industries.
The Concept Web revolutionizes web searches by assigning unique universal identifiers to each industry concept, applying the same ID to all synonymous terms across databases, grouping these terms into Concept Clouds, and matching a concept back to all of its synonyms during a search.
Code-N Analyze Box & Concept Display
As shown in this snippet of the Green Field Finder’s analysis of the identical search, “aspirin lowers blood pressure,” all 202 synonyms of “aspirin” will automatically be included (along with the 19 terms for “blood pressure” and seven terms for “lowers“).
Code-N Patent Results
The Green Field Finder retrieves results in seconds— claims involving all 26,866 combinations of the terms “aspirin,” “reduces,” “blood pressure” and their synonyms. No need to conduct and reconcile results from tens of thousands of individual searches! This is the way to accelerate complex research tasks. This is the way to tame Big Data.