Description
./viterbi.py input_hmm test_file output_file
I’m providing hmm.txt, which is in the following format (italics are replaced with values in the file):
state_num=number of states sym_num=number of POS tags
\init
tag P(tag | BOS) log10(P(tag | BOS))
…
\transition tag1 tag2 P(tag2 | tag1) log10(P(tag2 | tag1))
…
\emission tag word P(word | tag) log10(P(word | tag))
…
- If there is no transition probability from tag1 to tag2, then it is impossible to go from tag1 to tag2.
- If there is no transition probability from tag to word, then it is impossible for that tag to generate that word.
Each line of the test_file contains a single sentence.
The format of the output_file should be, one sentence per line: word1/tag1 word2/tag2 word3/tag3…
