An FNL parser as a proof of concept

To test and improve the FNL text-model, we implemented it in a functional grammar for an extended fragment of Cambridge English. The foundation for this FNLE grammar is: advanced feature tagging for both words and phrases, combined with the formal-functional structure of the production rules. Further, CALCULEMUS developed a pilot-parser for that fragment of Cambridge English.

This FNLE pilot-parser analyses an English input-text, sentence by sentence, in two steps:

  1. The parser produces the syntactic tree-structure (including all unified feature structures) of every consecutive text sentence,
  2. The parser produces the semantic (co-)referential ‘named entity analysis’, resulting in an identifying index for every determined phrase, with an accuracy tending to 100%.
 

Creative Rebels - Strategy, design en technology for your business.