Concept and Key content extraction

Modified on Tue, 28 Mar, 2023 at 9:39 PM

Concepts are the key terms of a patent, extracted using artificial intelligence (AI) technology, from the Fulltext of the publications. They reflect the semantic content of the publication. Using concepts in your searches can help you to be more exhaustive as you search semantically similar terms. Concepts are shown in the tab called "Concepts" in the right side panel. Concepts can also be charted which is discussed in more detail below.


For example, search "cellphone" in concepts and find patents referring to "mobile phones". The terms cellphone and mobile phones are conceptually similar. 

 


The Concept field has 3 values :
- the concept
- the percentage of relevance of this concept inside the document
- the frequency (number of apparition of this concept)

Low relevance concepts are often concepts with one or 2 apparitions, inside "less interesting" parts like in the middle of the description.
To the opposite, a concept (even with a frequency of 1) inside the object of the invention will have a better relevance score than other "isolate" concepts.
The list of concept is sorted by relevance score then by frequency (in case of equal score).

The relevance score is always a relative score. The most relevant concept has a score of 100, and all the others score are calculated from this one.
For its computing, these items are taken into account :
- position of the concept in the whole text
- position of the concept in its sentence
- frequency in the text and combination of the other scores for this concept

And at the end, the "raw" score is normalized, based on a 100% most relevant score of the document.


When exporting and when looking at the KWIC tab, concepts might be followed by two numbers in brackets. They refer to the importance of a concept (following a scale value from 0 to 100) in a document and to the frequency of the concept in a document.

For more info on the KWIC tab, please see the related article. 


The Key Content is extracted using linguistic technology from the fulltext and is made up of "Object of the Invention", "Advantages of the Invention Over Previous Art", and "Independent Claims". This tab is a summary of the invention and helps you to quickly review, understand and compare your hits.

 


Both Concepts and Key content are extracted from the full official text in English of the EP applications (without euro-PCT), the WO applications, patents from US, UK, CA, IN, AU, IL. It is also extracted from the English translations of patents from FR, CN, KR, JP and DE. Every concept is weighted according to the field in which it has been identified. The concepts are a result of the semantic content of the patent. 


Concepts can be charted in the analysis module to better understand your applications. In a saved analysis, you can also group or exclude concepts.


Learn more on saved analysis in the following article:  How to save an analysis and why.



Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article