metadata

license: other
license_name: topicnet
license_link: >-
  https://github.com/machine-intelligence-laboratory/TopicNet/blob/master/LICENSE.txt
configs:
  - config_name: bag-of-words
    default: true
    data_files:
      - split: train
        path: data/Brown_BOW.csv.gz
  - config_name: natural-order-of-words
    data_files:
      - split: train
        path: data/Brown_NOOW.csv.gz

Brown

The Brown Corpus was the first million-word electronic corpus of English, created in 1961 at Brown University. This corpus contains text from 500 sources, and the sources have been categorized by genre, such as news, editorial, and so on. 1.1 gives an example of each genre (for a complete list, see http://icame.uib.no/brown/bcm-los.html).

Language: English
Number of topics: 15
Number of articles: 500
Year: 1961

References

NLTK datasets: https://www.nltk.org/book/ch02.html.
Dataset site: http://korpus.uib.no/icame/brown.