dmytro_ushatenko/pages/topics/resources/README.md

86 lines
2.8 KiB
Markdown
Raw Normal View History

2020-10-01 14:05:57 +00:00
---
title: Slovenské jazykové zdroje
published: true
taxonomy:
category: [info]
2020-10-01 14:05:57 +00:00
tag: [annotation,ner,pos,question-answer,nlp]
author: Daniel Hladek
---
2020-03-02 13:49:10 +00:00
# Slovenské jazykové zdroje
### POS
2020-03-12 08:55:20 +00:00
[Multext East](http://nl.ijs.si/ME/) Anotovaný román George Orwell 1984 v 15 európskych jazykoch
2020-03-02 13:49:10 +00:00
### NER
- Learning multilingual named entity recognition from Wikipedia- WIKI Ner?
- Cross-lingual Name Tagging and Linking for 282 Languages - NER anotácia aj slovenskej Wikipédie podľa anglickej
- https://drive.google.com/drive/folders/1bkK6ly_awxe9IgAKL16VVvCtjcYcDSw8
- https://elisa-ie.github.io/wikiann/
2020-03-02 13:49:10 +00:00
### Parsing-POS
[Slovak Dependency Treebank](https://lindat.mff.cuni.cz/repository/xmlui/handle/11234/1-1822)
https://github.com/UniversalDependencies/UD_Slovak-SNK
[Artificial Treebank with Ellipsis](https://lindat.mff.cuni.cz/repository/xmlui/handle/11234/1-2616)
### Wordnet
[Slovak Word Net](https://korpus.sk/WordNet.html)
### Parallel Corpus
2020-03-02 13:49:10 +00:00
Europarlament
[Czech-Slovak Parallel Corpus](https://lindat.mff.cuni.cz/repository/xmlui/handle/11858/00-097C-0000-0006-AADF-0)
[English-Slovak Parallel Corpus](https://lindat.mff.cuni.cz/repository/xmlui/handle/11858/00-097C-0000-0006-AAE0-A)
[Multext East](http://nl.ijs.si/ME/)
2020-03-12 08:55:20 +00:00
### Sentiment
2020-03-02 13:49:10 +00:00
[Twitter sentiment for 15 European languages](https://www.clarin.si/repository/xmlui/handle/11356/1054)
### Web
- [Aranea](http://ucts.uniba.sk/aranea_about/)
- [SkTenTen](https://www.sketchengine.eu/sktenten-slovak-corpus/) automaticky POS anotovaný, prístup cez web rozhranie
- [CommonCrawl](https://commoncrawl.org/2020/03/february-2020-crawl-archive-now-available/) Obsahuje aj slovenské dáta?
- [Oscar](https://traces1.inria.fr/oscar/) klasifikácia a deduplikácia dát z COmmonCrawl, aj pre slovenčinu (4.5 GB dedub, 665M slov dedup.)
2020-03-02 13:49:10 +00:00
2020-03-12 08:55:20 +00:00
### Wikipedia
[Wikipedia vo formáte JSON Elasticsearch Bulk](https://dumps.wikimedia.org/other/cirrussearch/current/)
2020-03-02 13:49:10 +00:00
### Word Embedding
- [FastText Word Embedding from Common Crawl](https://fasttext.cc/docs/en/crawl-vectors.html)
- [FastText Word Embedding from Wikipedia](https://fasttext.cc/docs/en/pretrained-vectors.html)
2020-03-02 13:49:10 +00:00
### Databázy zdrojov
https://github.com/slovak-nlp/resources
2020-03-02 13:49:10 +00:00
https://www.clarin.eu/portal
https://www.clarin.eu/resource-families/manually-annotated-corpora
http://www.meta-share.org/
https://korpus.sk/res.html
Slovak Stemming https://www.mediawiki.org/wiki/User:TJones_(WMF)/Notes/Slovak_Stemmer_Analysis
### Tools
- [Spacy](https://spacy.io/), tokenizer, stopwords, custom model
- [Slovak Lexer](https://github.com/hladek/slovak-lexer) / tokenizer
- [Slovak Elasticsearch](https://github.com/essential-data/elasticsearch-sk) - stopwords, stemmer
- [Slovak Hunspell](https://github.com/essential-data/hunspell-sk) - stemmer, spelling