Using parallel corpora for automatic transfer of FrameNet annotation

Richard Johansson, Pierre Nugues

Research output: Chapter in Book/Report/Conference proceedingPaper in conference proceedingpeer-review

Abstract

We present a method for producing FrameNet-annotated text for non-English text. The method uses sentence-aligned corpora and transfers target words and frame elements using a word aligner.
The system was tested on an English-Spanish parallel corpus. On the task of target word projection, the system had a precision of 69% and a recall of 59%. For sentences
with non-empty targets, it had a precision of 84% and a recall of 81% on the task of transferring frame elements. The approximate precision of the complete frame element bracketing of Spanish text was around 64%.
Original languageEnglish
Title of host publicationProceedings of the 1st ROMANCE FrameNet Workshop
EditorsVincenzo Pallotta, Dan Tufiş
Number of pages7
Publication statusPublished - 2005
Event1st ROMANCE FrameNet Workshop - Cluj-Napoca, Romania
Duration: 2005 Jul 262005 Jul 28

Conference

Conference1st ROMANCE FrameNet Workshop
Country/TerritoryRomania
CityCluj-Napoca
Period2005/07/262005/07/28

Subject classification (UKÄ)

  • Computer Science

Free keywords

  • Natural language processing
  • frame semantics
  • parallel corpora

Fingerprint

Dive into the research topics of 'Using parallel corpora for automatic transfer of FrameNet annotation'. Together they form a unique fingerprint.

Cite this