Hidden Markov Model Part of Speech Tagger using the Viterbi algorithm
Requirments: Python 3.x
Training:
Invoked with ./train_set.py -f <path to training file>
Other arguments can be seen through ./train_set.py --help
Viterbi:
Invoked with ./viterbi.py
With no arguments it assumes there is a file in the invoking directory called 'WSJ_24.words'
You can specify a sentence file with the '-s' argument as in ./viterbi.py -s <path_to_sentence_file> All sentence files are assumed to be a word on each line and sentences seperated by a new line
Other arguments can be seen through ./viterbi.py --help
Note: This program expects two json files in the same directory: trainingDataPOS.json trainingDataWORD.json