Using the below list, I need help creating some HMM models and predicting which model works best.

list $=[[("$the$",$ "det"$),("$further$",$ "adv"$),("$which$",$ "det"$),("$constructed$",$ "verb"$),("$report$",$ "noun"$)],[("$been$",$ "verb"$),("$judge$",$ "noun"$),("$pastor$",$ "noun"$),("$the$",$ "det"$),("$in$",$ "adj"$),("$was$",$ "verb"$)],[("$that$",$ "adp"$),("$to$","$prt$"),("$and$",$ "conj"$),("$johnson$",$ "noun"$),("$it$",$ "pron"$),("$jury$",$ "noun$)]]$

Use sklearn to split the list above.

Use NLTK lidestoneProbDist, MLEProbDist, and ELEProbDist to create hmm models and train$\_$supervised method from HiddenMarkovModel trainer to pass the training set.

Use best$\_$path function to predict the tags from the test dataset from above.