Reuters Corpus, Volume 1

registration: 2007/10/18 06:33:05, last modified: 2007/11/11 04:53:46
title
  • Reuters Corpus, Volume 1
title.abbreviation
title.alternative
  • RCV1
creator
  • developer Reuters, Ltd.
subject
  • test collection for text categorization research
subject.linguisticField
  • computational_linguistics
subject.monoMultilingual
  • monolingual
subject.resourceSubject
  • corpus
description
  • RCV1 is a large collection of news stories containing about 810,000 Reuters, English Language News stories. It is significantly larger than the older Reuters-21578 collection heavily used in the text classification community.
description.price *This metadata is only as a guide
  • (probably)free for research purpose
description.language
  • English
description.inputDevice
description.inputEnvironment
description.speakingStyle
description.speechMode
description.samplingRate
description.additionalData
publisher
  • the National Institute of Science and Technology(reuters-request@nist.gov)
contributor
contributor.attribute.author.motherTongue
  • native Reuters journalists
contributor.attribute.author.dialect
  • standard_dialect
contributor.attribute.author.level
  • professional
contributor.attribute.author.age
  • adult
contributor.attribute.author.gender
  • male
  • female
contributor.attribute.speaker.motherTongue
contributor.attribute.speaker.dialect
contributor.attribute.speaker.level
contributor.attribute.speaker.age
contributor.attribute.speaker.gender
contributor.attribute.speaker.number
date.created
  • 1996-00-00 1997-00-00
date.issued
  • 2000-11-03
date.modified
type
  • Text
type.discourseType
  • report
type.linguisticType
  • primary_text
type.purpose
  • analysis
  • developing_technologies
  • developing_technologies For use in research and development of natural language processing, information retrieval, and machine learning systems
type.style
  • written news paper
type.form
  • unfixed
type.sentence
  • short
  • long
type.annotation
  • annotated
type.annotationSample
  • CATEGORY CODE (assigned to each story):
    Topics
    Industries
    Regions

    TOPIC CODES (major subjects of a story):
    CCAT (Corporate/Industrial)
    ECAT (Economics)
    GCAT (Government/Social)
    MCAT (Markets)

    INDUSTRY CODE (type of businesses discussed in the story): Grouped in 10 subgroups, such as I2 (METALS AND MINERALS) and I5 (CONSTRUCTION).

    REGION CODES (both geographic locations and economic/political groupings. No subgroup)

    David D. Lewis, "RCV1: A New Benchmark Collection for Text Categorization Research"
    http://jmlr.csail.mit.edu/papers/volume5/lewis04a/lewis04a.pdf
type.sample


  • USA: Tylan stock jumps; weighs sale of company.
    Tylan stock jumps; weighs sale of company.
    SAN DIEGO

    The stock of Tylan General Inc. jumped Tuesday after the maker of process-management equipment said it is exploring the sale of the company and added that it has already received some inquiries from potential buyers.


    Tylan was up $2.50 to $12.75 in early trading on the Nasdaq market.


    The company said it has set up a committee of directors to oversee the sale and that Goldman, Sachs & Co. has been retained as its financial adviser.



    (c) Reuters Limited 1996























    David D. Lewis, "RCV1: A New Benchmark Collection for Text Categorization Research"
    http://jmlr.csail.mit.edu/papers/volume5/lewis04a/lewis04a.pdf
format.extent
  • about 810,000 news stories on two CDs. Requires about 2.5 GB for storage of the uncompressed files.
format.medium
  • CD
format.encoding
format.markup
  • XML
format.functionality
identifier
identifier.doi
identifier.islrn
source
  • Reuters News stories
language
  • English
language.area
  • Europe
relation
relation.utilization *This metadata is automatically extracted
  • To illustrate our hypothesis, we performed a series of experiments on named entity recognition using a set of English data from the Reuters corpus.
  • We built a large dependency database for English based on an automatic parse of the BNC, and Reuters (sports and finances sections).
  • In order to make efficient searches in the English parsed corpora we created a database for each corpus (BNC, Reuters_sports, Reuters_finances), where each tuple represents a dependency syntactic relation between a verb and a dependent (head of a noun or prepositional phrase).
  • Interest Changes To evaluate Nootropia's ability to adapt to a variety of interest changes, we synthesised virtual users and simulated changes in their interests using combinations of RCV topics.
coverage.temporal
  • 1996/08/20-1997/08/19
coverage.spacial
rights