tagger

Documentation

Overview

Package tagger provides a Hidden Markov Model part-of-speech tagger.

Index

• type HMMTagger
  • func NewHMMTagger(model model.Model, wordHandler words.WordHandler, ...) HMMTagger
  • func (t HMMTagger) Tag(sentence []string) Trellis
• type Trellis
  • func (t Trellis) Tags() ([]string, float64)

Types

type HMMTagger

type HMMTagger struct {
	// contains filtered or unexported fields
}

HMMTagger implement a Hidden Markov Model (HMM) part-of-speech tagger.

func NewHMMTagger

func NewHMMTagger(model model.Model, wordHandler words.WordHandler,
	trigramModel trigrams.TrigramModel, beamFactor float64) HMMTagger

NewHMMTagger constructs a new tagger from the given data model, word handler and trigram model. The beam factor specifies how aggressively the search space should be pruned. For instance, a beam factor of 1000 will exclude all paths that are 1000 times less probable than the most probable path.

func (HMMTagger) Tag

func (t HMMTagger) Tag(sentence []string) Trellis

Tag tags a sentence.

type Trellis

type Trellis struct {
	// contains filtered or unexported fields
}

A Trellis is used during HMM tagging to store possible analyses.

func (Trellis) Tags

func (t Trellis) Tags() ([]string, float64)

Tags returns the most likely part-of-speech tag sequence in the Trellis.