The content of individual sentences is compared to the content
of the whole document using InfoCodex' unique similarity
measure. The sentences with the closest similarities compared to
overall content are extracted to form an abstract. This procedure guarantees
that the essential facts are featured in the abstract, which represents
the core contents of a document.
The abstracts are factual and standardized in size. They
facilitate the content viewing and targeted finding of relevant content
(-> substantial time savings for the user).