Inventions (patent families) are ordered by default by relevance.
The relevancy of inventions is mainly based on the following indicators :
- Number of occurrences of the searched concept
- Position of the concept within the document
- The concepts: The concepts are extracted from the full text in official English of EP applications (excluding Euro-PCT) since 1988, PCT applications since mid-2001, patents granted US from 1971 to 2000, US applications since 15/03/2001 and GB applications since 1979, non-extended RF applications, CN utility models and CN applications.PCTs (not published in Korean) translated into machine English are also used to extract the concepts. Some older documents, up to 1980 for EPs and 2000 for WOs are present. First, nominal sentence are identified within the full text, then they are normalized according to syntactic and semantic rules. Finally, each concept is weighted according to the field in which it has been identified and its occurenc. The concepts reflects the semantic content of the patent.