README
¶
Citar - Trigram HMM part-of-speech tagger
Introduction
Citar is a part-of-speech tagger, based on a trigram Hidden Markov Model (HMM). It (partly) implements the ideas set forth in [1]. It can be used as a set of stand-alone programs and or from Go.
C++ Citar
The C++ version of Citar can be found is still available. Active maintenance will proceed in Go. The choice to port Citar from C++ to Go was not made lightly. However, I believe that switching to Go will ease maintenance, lower the barrier for contributions, and improve cross-platform support. Moreover, recent version of Go have made it easier to call Go code from C.
[1] TnT - a statistical part-of-speech tagger, Thorsten Brants, 2000
Documentation
¶
Overview ¶
Package citar provides an HMM part-of-speech tagger.
citar is a package that provides an Hidden Markov Model part-of-speech tagger. The tagger can be used as a set of command-line utilities or as a Go package for integration in Go applications.
The architecture of this tagger is inspired by Thorsten Brants' TnT tagger: TnT: A Statistical Part-of-Speech Tagger, Thorsten Brants, Proceedings of the sixth conference on Applied natural language processing, ANLC '00
Source Files
Directories
¶
| Path | Synopsis |
|---|---|
|
cmd
|
|
|
Package model provides the tagger's data model.
|
Package model provides the tagger's data model. |
|
Package tagger provides a Hidden Markov Model part-of-speech tagger.
|
Package tagger provides a Hidden Markov Model part-of-speech tagger. |
|
Package trigrams provides trigram models.
|
Package trigrams provides trigram models. |
|
Package words provides methods to estimate (word) emission probabilities.
|
Package words provides methods to estimate (word) emission probabilities. |