Generating a controlled natural language grammar from example business rules

From Master Projects
Jump to: navigation, search

About Generating a controlled natural language grammar from example business rules

  • This project has not yet been fulfilled.
  • This project fits in the following Bachelor programs: {{#arraymap:|, |xXx|bachelorproject within::xXx|,}}
  • This project fits in the following masterareas: {{#arraymap:AI and Communication,

Computer Science and Communication, Human Ambience, Information Sciences, Information and Communication Technology, Internet and Web Technology, Knowledge Technology and Intelligent Internet Applications,|, |xXx|project within::xXx|,}}


This project would be performed in collaboration with LibRT:

Problem: In natural language we may express the same meaning in multiple ways. That is an adventage when writing a novel, but a disadventage when writing operational guidance as business rules. We want authors to express the same kind of knowledge in a similar way. But our rule authors can not write a formal grammar. They are sales professionals, accountants, medical experts, brokers, traffic engineers – they all have to deal with rules.

Solution: We have a business rules editor named RuleXpress that allows our subject matter experts to autor business rules supported by a structured vocabulary (SBVR-based) and a controled natural language. The CNL is defined as a context free grammar in XML.

Method: Instead of defining a context free grammar our user would like to give a set of example business rules and have the system generate a CNL pattern that matches the sentences. The algorithm may use information from a part of speech tagger to distinquish between verbs / nouns etc. It would be cool if the method works for multiple natural languages.