A POS tagging project uses HMM and Viterbi written in Python by 190000096. This project investigates whether the use of Witten-Bell Smoothing and UNK tags will improve the tagging accuracy in English and other languages.
This project includes three .py files: 1.smooth.py : This smoothing program investigates whether the use of Witten-Bell Smoothing will improve the accuracy. 2.tagging.py: This tagging program investigates whether the use of UNK tags will improve the accuracy in English. It may takes about half an hour to output. 3.tagging_for_portuguese.py : This tagging program investigates whether the use of UNK tags will improve the accuracy in Portuguese language.
1.Python 3.8 2.NLTK Package 3.Texttable package
$ python3 code/smooth.py
$ python3 code/tagging.py
$ python3 code/tagging_for_portuguese.py