CYK (Cocke-Younger-Kasami) Grammar Rules

1.3k views Asked by At

I am interested in Natural Language Parsing and have written a Brill Part of Speech Tagger, and would like to enhance it by combining it with a a POS tagger based on grammar rules. Is anyone aware of open source ruleset files for English anywhere? I am particularly interested in anything related to the CYK (Cocke-Younger-Kasami) algorithm, C# especially. Thanks.

1

There are 1 answers

0
Yasen On

Have you checked GATE?

They seem to have a Brill-based POS tagger with lexicons and rules. It's implemented in Java so you should probably understand the code easily. It is available in the same project at sourceforge. Here is their POS tagger documentation.

Hope this helps.

Off topic, lectures from Michael Collins on POS tagging and theory behind it.