Published August 25, 2021 | Version v1

Reuters90

Authors/Creators

Description

Reuters90 (REUT)} is a classical text dataset, composed by news articles collected and annotated by Carnegie Group, Inc. and Reuters, Ltd. We consider here a set of 13, 327 articles, classified into 90 categories.

The files:
texts.txt: Document set (text). One per line.
score.txt: Document class whose index is associated with texts.txt
split_<k>.pkl:  pandas DataFrame with k-cross validation partition.

Files

reut90.zip

Files (102.0 MB)

Name Size Download all
md5:42acf8b07974bdfcff3d09bfc2c1903c
89.1 MB Preview Download
md5:3b88eeebc416e9bb32edc4c7b2a2d51c
39.6 kB Preview Download
md5:68c8dfc5d9ba1d3922f405b75546b61b
398.5 kB Download
md5:964c3507c883fc701a1f30d26afe2e99
199.7 kB Download
md5:7a625b610169142fddcc917f43cfec62
12.3 MB Preview Download

Additional details