The project does POS tagging on tagged brown corpus using Hidden Markov Model (Viterbi Algorithm) and CRF (Conditional Random Fields)
Accuracy of Viterbi Algorithm:
F1-Scores (Number of Test Sentences Used = 10) :
Precision - 0.935
Recall - 0.925
F1-Score - 0.925
Support Count- 239
Accuracy of CRF(Conditional Random Fields):
F1-Scores (Number of Test Sentences Used = 10) :
Precision - 0.961
Recall - 0.958
F1-Score - 0.957
Support Count- 239