Skip to content

52nlp/twitter_nlp

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

55 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Twitter NLP Tools

Authors: Alan Ritter, Sam Clark

contact: aritter@cs.washington.edu

Example Usage:

export TWITTER_NLP=./
cat test.1k.txt | python python/ner/extractEntities2.py

note: this takes a minute or so to read in models from files

To include classification, simply add the --classify switch:

cat test.1k.txt | python python/ner/extractEntities2.py --classify

For higher quality, but slower results, optionally include features based on POS and chunk tags (chunk tags require POS)

cat test.1k.txt | python python/ner/extractEntities2.py --classify --pos
cat test.1k.txt | python python/ner/extractEntities2.py --classify --pos --chunk

Also has the ability to include event tags (requires POS):

cat test.1k.txt | python python/ner/extractEntities2.py --classify --pos --event

Output:

The output contains the tokenized and tagged words separated by spaces with tags separated by forward slash '/' Example output:

The/B-movie/DT/B-NP/O Town/I-movie/NNP/I-NP/O might/O/MD/B-VP/O be/O/VB/I-VP/O one/O/CD/B-NP/O of/O/IN/B-PP/O the/O/DT/B-NP/O best/O/JJS/I-NP/O movies/O/NNS/I-NP/O I/O/PRP/B-NP/O have/O/VBP/B-VP/O seen/O/VBN/I-VP/O all/O/DT/B-NP/O year/O/NN/I-NP/O ./O/./O/O So/O/RB/O/O ,/O/,/O/O so/O/RB/B-ADJP/O good/O/JJ/I-ADJP/O ./O/./O/O And/O/CC/O/O don't/O/NN/B-NP/O worry/O/NN/I-NP/O Ben/B-person/NNP/I-NP/O ,/O/,/O/O we/O/PRP/B-NP/O already/O/RB/B-ADVP/O forgave/O/VBP/B-VP/B-EVENT you/O/PRP/B-NP/O for/O/IN/B-PP/O Gigli/B-movie/NNP/B-NP/O ./O/./O/O Really/O/RB/B-INTJ/O ./O/./I-INTJ/O

Looking at just one word:

The/B-movie/DT/B-NP/O

The fields are as follows:

Word: The
Entity: B-movie Begins a named entity of type "movie"
Chunk: B-NP Begins a noun phrase
Event: O Not part of an event phrase

The BIO encoding is used for encoding phrases (Named Entities, event phrases, and chunks), for example:

The/B-movie Town/I-movie might/O ...

Indicates that the word "The" begins a named entity of type movie, "Town" continues that entity, and "might" is outside of an entity mention. For more details see: http://nltk.org/book/ch07.html

Requirements:

  1. Linux
  2. Libraries and executables can be compiled with build.sh

Relevant papers:

@inproceedings{Ritter11,
  author = {Ritter, Alan and Clark, Sam and Mausam and Etzioni, Oren},
  title = {Named Entity Recognition in Tweets: An Experimental Study},
  booktitle = {EMNLP},
  year = {2011}
}

@inproceedings{Ritter12,
  author = {Ritter, Alan and Mausam and Etzioni, Oren and Clark, Sam},
  title = {Open Domain Event Extraction from Twitter},
  booktitle = {KDD},
  year = {2012}
}

Demo:

statuscalendar.com

Acknowledgments (bug fixes, etc...):

Junming Sui

Ming-Wei Chang

Tuan Anh Hoang Vu

sumant81

Yiye Ruan

Lu Wang

About

Twitter NLP Tools

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • HTML 84.1%
  • Java 12.0%
  • Nu 2.0%
  • C 0.8%
  • Python 0.5%
  • Perl 0.4%
  • Other 0.2%