This is a part-of-speech (POS) tagging project which I did when I took CSCI544 Applied Natural Language Processing in Fall 2021 at USC.
For this project, I built an HMM model for POS tagging. I tried two decoding methods: greedy and viterbi.
The accuracy of the greedy decoding algorithm with HMM is 92.67%, while the accuracy of the viterbi decoding algorithm with HMM is 94.36%.