Knowledge Yielding Ontologies
for Transition-based Organization
  • Increase font size
  • Default font size
  • Decrease font size

Kaf annotator


Stand-alone application for annotating KAF files with any set of tags to any level. This annotator is used to create gold-standard data for evaluating the Kybot output.

The program allows you to load any KAF file and any tag set to annotate up to 8 levels for each word token. It also allows you to assign synsets from different wordnets available on the KYOTO server. The tag set can be organized hierarchically and can be modified while tagging. The tags are stored at the word token level, but you can assign a tag to all the word tokens belonging to a term, a chunk or a sentence. The tags are saved in a separate file and can be loaded on top of the KAF file.

The interface provides various possibilities to sort the words or codes in any order and reconstruct the word order of the text. Likewise, you can apply in context and out of context tagging at any time. Finally, a lexicon is build during the tagging. This lexicon can be used to assign the most frequent tag to any token.

A tag file can be  converted to a triplet file as defined for the KYOTO fact evaluation. The triplet conversion requires a configuration file to specify at what level the event tags are found and what levels are used to indicate the scope of the event. The triplet file generated from a tag file can be used as a gold standard for evaluation.

There is extensive documentation included in the package.

 To run the Kaf Annotator, unpack the zip file in any folder and double click the .bat file in Windows or run the .sh file from Linux/Unix. The program requires Java1.6 to run.

Owner: VU University Amsterdam

License: GPLv3


Source code:

To compile run: mvn install



ICT-211423 - 2008 © Kyoto Consortium