Math.HiddenMarkovModel.Pattern

This module provides a simple way to train the transition matrix and initial probability vector using simple patterns of state sequences.

You may create a trained model using semigroup combinators like this:

 let a = atom $ HMM.state 0
     b = atom $ HMM.state 1
     distr =
        Distr.DiscreteTrained $ Map.fromList $
        ('a', Vector.fromList [1,2]) :
        ('b', Vector.fromList [4,3]) :
        ('c', Vector.fromList [0,1]) :
        []
 in  finish 2 distr $ replicate 5 $ replicate 10 a <> replicate 20 b

Documentation

data T prob

Instances

Field prob => Semigroup (T prob)

atom :: Container Vector prob => State -> T prob

append :: Container Vector prob => T prob -> T prob -> T prob

replicate :: Container Vector prob => Int -> T prob -> T prob

finish :: Container Vector prob => Int -> tdistr -> T prob -> Trained tdistr prob