Natural Language Processing Resources
From AGIRI.org
Natural language procesing is curently understood to require several components. These include dictionaries (or NLP databases or netoworks), parsers, and ontologies. The last component, ontologies, are used to provide a semantic foundation for disambiguating ambiguous sentances. The distinction between a "dictionary" used in the parsing, and an ontology, can be blurry.
Contents |
[edit]
NLP databases
Some important NLP databases are:
- WordNet
- FrameNet
- Something called Preposition WordNet was constructed at Novamente LLC; if you want to use it, contact ben@goertzel.org
[edit]
Open source parsers
- NLTK, the Natural Language Toolkit. Documentation includes a book, multiple articles. Features integration into WordNet. Written in python.
- Link Grammar Parser, from Carnegie-Mellon. A parser for the English language, based on "link grammar", a novel theory of natural language syntax. Written in C, with a BSD-like license that is compatible wth the GPL.
- GATE, General Architecture for Text Engineering. Old, big, well-documented, with many features and bits and pieces, including dialog processing and NL generation. Written in Java, GPL license.
[edit]
Commercial and closed source parsers
Commercial or closed-source research parsers include:
- Cyc NL subsystem, from Cyc corporation.
- MiniPar, a parser created by Dekang Lin. Closed-source, free for non-commercial use.
[edit]
