HW 10: HIDDEN MARKOV MODELS

1. Following the coin-toss example in lecture no. 22, train all three models on the following data sequence:
   
   
   "HTHTHHTHTHHHTHTHHHTTTTHTHTHHHTHTHTHHHTHTHTHHHHTHTHTHHHTHTHTHTHTH HTHTHTHTHHHHTHTHTHHHTHTHTHHHTHTHTHTHHTHTHTHTHTHTHTHTHHTHHHHHHTHTHT"
   
   using the Baum Welch training algorithm. Which model best matches the data?
   
   
2. Using your best model, generate 1000 coin tosses. Train the same three models on this data. Compare these models to the model that generated the data. Explain.