Implemented a NER Tagging algorithm with Hidden Markov Model:
- Implementing Learning (Counting) and Decoding (Viterbi)
- Training the system on Wikipedia ENG and ITA
- Evaluate the system, using different strategies of smoothing for ENG and ITA
- Evaluate against an easy baseline and a difficult one
96.32% accuracy for IT and 97.44% for ENG.
For more details see the report.