Additional resources for automatic keyphrase extraction and term assignment

• Additional resources for automatic keyphrase extraction and term assignment
• FAO-780 data set for term assignment
• Other data sets on the web
• Vocabularies
• Competing systems and demos

Apart from the three collections listed in the MultiplyIndexedData, Maui was also tested on several term assignment collections, where each document had a single set of manually assigned topics.

FAO-780 data set for term assignment

• Download FAO-780

This data set contains 780 documents with terms from Agrovoc assigned by professionals. When using this data set, please cite Medelyan (2009) or Medelyan and Witten (2008), see Publications.

Other data sets on the web

NUS Keyphrase Corpus can be used for training and testing keyphrase extaction and tagging. To use it with Maui, the data first needs to be converted into the required format.

SemEval-2010 Keyphrase Extraction will soon publish their data for the participants of the shared task. This data set will be similar to the NUS Keyphrase corpus.

See also: MultiplyIndexedData

Vocabularies

For term assignment, a number of vocabularies in SKOS format are available on the web:

• Library of Congress Subject Headings LSCH
• Medical Subject Headings thesaurus MeSH

Please note: Maui user Amrita recommends to use HIVE to convert original MeSH vocabulary into SKOS.

• FAO’s agricultural thesaurus Agrovoc: general info and download site.
• List of other SKOS thesauri at FAO
• DESY’s High Energy Physics HEP thesaurus
• W3C’s list of SKOS thesauri

Competing systems and demos

The topic indexing blog provides a list of tools for automatic keyphrase extraction, terminology extraction, tagging and other tasks.