Please build your character$-$gram $($char$-$gram$)$ language models for the given training set. Please assume that you experiment will only have the following characters

$[$a$,$ b$,$ c$,$ d$,$ f$,$ h$]$ that exists in the training set.

Training set:

a a b c h d b f

b c h h a d f f a h b

b b a a h h c c h d d f a b f f h c c d f f h

h h f c f c a a c c d d d

Test set:

a b c d f

a b b c c d

Task:

Build char$-$unigram language model,

Build char$-$bigram language model,

Compute joint probability for the given test set using char$-$unigram model.

Compute perplexity of your models $($char$-$unigram, char$-$bigram$)$ and compare which model is better.