20NewsGroup

Electronic messages posted in 20 newsgroups.
Source: http://qwone.com/~jason/20Newsgroups/

11,314 messages
56,126 words

Files:
* biadjacency.tsv : list of edges (message, word)
* names_col.tsv : words
* labels.tsv : labels of messages (7 categories)
* names_labels.tsv : names of these labels
* labels_hierarchy.tsv : hierarchical labels of messages (20 newsgroups)
* names_labels_hierarchy.tsv : names of these labels

Note: The 7 categories subsets of the 20 newsgroups:

computers -> comp.* (5)
autos -> rec.autos, rec.motorcycles
sport -> rec.sport.baseball, rec.sport.hockey
science -> sci.* (4)
forsale -> misc.forsale
politics -> talk.politics.* (3)
religion -> alt.atheism, soc.religion.christian, talk.religion.misc