Pos-taggers

Part-of-speech taggers for Kazakh language. Used PoS tagger models:

• Trigrams'n'Tags(TnT) from python's nltk package.
• RDRPOSTagger
• Conditional Random Fields (CRFs)

Used datasets:

• Universal Dependencies(UD)
• Kazakh Language Corpora(KLC)

Installation

Run conda create -n myenv --file package-list.txt

Calculate accuracy of RDRPOSTagger

For KLC

Run python Eval.py full-path-to/klc.raw_test.TAGGED full-path-to/klc.test

For UD

Run python Eval.py full-path-to/kazcorpus_raw_test.TAGGED full-path-to/kazcorpus_test

For more details see http://rdrpostagger.sourceforge.net/