Institut

Studium

Forschung


 

Download

TIGER 2.2.doc (April 2018)

Version 2.2.doc is a new set of annotations for the Release 2.2. It includes a full mapping of sentences to documents, as well as additional sentence-level and document-level annotations.

TIGER Corpus Release 2.2 (July 2012)

Version 2.2. is a correted version of Release 2.1. provided by Wolfgang Seeker.

TIGER Corpus Release 2.1 (August 2007)

The lastest TIGER Corpus distribution is bundled in one ZIP file. This file comprises:

  • TIGER Corpus (Negra Export format)
  • TIGER Corpus (TIGER-XML format)
  • Documentation

Please download the following file:

TIGER Corpus Release 2 (December 2005)

You can still download the TIGER Corpus Release 2 (28 MB). It is bundled in the same way as Release 2.1.

TIGER Corpus Release 1 (July 2003)

Also, you can still download the TIGER Corpus Release 1. Its distribution is also bundled in one ZIP file, which comprises:

  • TIGER Corpus (Negra Export format)
  • TIGER Corpus (TIGER-XML format)
  • TIGER Corpus (PennTreebank format)
  • Documentation

Please download the following file:

TIGER data sets for CoNLL-X shared task (May 2006)

The organizers of the CoNLL-X shared task on multilingual dependency parsing prepared the TIGER data for the task. You can download all relevant files as one BZIP2 tar ball:

TIGER 10.000 MOD Bank

The data includes the first 10,000 sentences from the TIGER Corpus (Release 2.1), where the original POS tags have been replaced by new tags that provide a more fine-grained analysis of modification in German. The sentences have been automatically tagged with the new tag distinctions and have been checked by one expert annotator.
For more details on the annotation scheme please refer to the annotation guidelines.

Please download the following file:

The TIGER 10.000 MOD Bank is provided by Ines Rehbein (irehbein AT uni-potsdam DOT de) and Hagen Hirschmann (hirschhx AT hu-berlin DOT de).

TiGer Dependency Bank

The TiGer Dependency Bank (TiGer DB) was created for sentences 8,001 through 10,000 of the TIGER Corpus as a dependency-based gold standard for German parsers. Its annotation differs somewhat from the TIGER Treebank annotation in that it allows for reentrancies and distinguishes more types of dependencies than the TIGER Treebank. Its closest relative is probably the PARC 700 Dependency Bank. The TiGer DB ZIP bundle comprises:

  • TiGer Dependency Bank
  • Documentation
  • Emacs library tigerDB-display.el for the convenient check and display of TiGer DB structures, similar to feature-display.el, which comes with the PARC 700 DB.

Please download the following file:

If you notice annotation errors or inconsistencies in the TiGer DB, please report them to Aoife Cahill (aoife DOT cahill AT ims DOT uni-stuttgart DOT de).

TIGER 700 RMRS Bank

The TIGER 700 RMRS Bank consists of Robust Minimal Recursion Semantics (RMRS) annotations of 700 sentences from the TiGer DB. It is joint work by Kathrin Spreyer and Anette Frank, then of Saarland University. The TIGER 700 RMRS Bank ZIP bundle comprises:

  • TIGER 700 RMRS Bank
  • Documentation
  • Tools for displaying TIGER 700 RMRS Bank structures etc.

Please download the following file:

For feedbank concerning the TIGER 700 RMRS Bank, please contact Kathrin Spreyer (spreyer AT uni-potsdam DOT de).

TIGER Dependency Triples

The dependency triple structures that Manuel Kountz extracted fully automatically from almost the entire TIGER Treebank as part of his Studienarbeit are formatted in the same way as the TiGer DB, but the structures contain reentrancies only to the extent as they are annotated in the treebank by means of secondary edges, and the dependencies encoded are as similar as possible to the ones in the TIGER Treebank. The most important structural difference can be found in the annotation of analytic tenses and passives, which receive a flat analysis in the dependency triples, whereas they are associated with a layered structure in the treebank. The TIGER Dependency Triples ZIP bundle comprises:

  • TIGER Dependency Triples
  • Documentation (in German)

Please download the following file:

 

Visitor Counter by Digits
hits since September 2007 according to
wc