Skip to main content

Hannover Tagger: Morphological Analysis and POS Tagging

Project description

HanTa - The Hanover Tagger

HanTa is a pure Python package for lemmatization and POS tagging of Dutch, English and German sentences. The approach is to some extent language indpendent and language models for more langauges will be added in future.

Lemmatization and POS tagging are based on the morphological analysis of a word. The morphological analysis is done by an Hidden Markov Model that tries to find the best sequence of morphemes underlying each word.

Usage

First a model has to be loaded:

from HanTa import HanoverTagger as ht

tagger_de = ht.HanoverTagger('morphmodel_ger.pgz')
tagger_nl = ht.HanoverTagger('morphmodel_dutch.pgz')
tagger_en = ht.HanoverTagger('morphmodel_en.pgz')

Now we have three methods to anaylze words and sentences:

tagger_en.tag_word('eating')

will give a list of all possible parts of speech (PoS) for the word eating together with a probability score.

tagger_en.tag_word('eating')

The function analyze gives the most likely PoS and the lemma (VBG and eat in the exmaple below).

tagger_en.analyze('unhappiest')

Using various optional parameters we can get more information like e.g. a list of morphemes:

tagger_nl.analyze('huishoudhulpje',taglevel=3)

The last call producses the following output:

('huishoudhulp', [('huis', 'N(soort,onz)'), ('houd', 'WW'), ('hulp', 'N(soort,zijd)'), ('je', 'SUF_DIM')], 'N(soort,ev,dim,onz,stan)')

The package also contains a simple trigram based PoS tagger, that uses the probabilities from the morphological analysis for unknown words (and infrequent words from he training data).

import nltk
from pprint import pprint

sent = "Die Europawahl in den Niederlanden findet immer donnerstags statt."

words = nltk.word_tokenize(sent)
lemmata = tagger_de.tag_sent(words)
pprint(lemmata)

Further reading

For more information refer to the following resources:

  • The main documentation: The Hanover Tagger (Version 1.1.0) - Lemmatization, Morphological Analysis and POS Tagging in Python. Hannover, 2023 Online Available
  • Demo.ipynb on GitHub (https://github.com/wartaal/HanTa)
  • Original publication: Christian Wartena. A probabilistic morphology model for German lemmatization. In Proceedings of the 15th Conference on Natural Language Processing (KONVENS 2019): Long Papers, pages 40–49, Erlangen, Germany, 2019. German Society for Computational Linguistics & Language Technology. Online Available

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

HanTa-1.1.1.tar.gz (15.0 MB view hashes)

Uploaded Source

Built Distribution

HanTa-1.1.1-py3-none-any.whl (15.0 MB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page