Published February 7, 2020 | Version V.2020

Negation handling for Amharic sentiment classification

Authors/Creators

  • 1. Addis Ababa University

Description

All four data sets are labeled for either "positive" or "negative". Tow of the data sets : ebc_facebook_news_only_Amharic_texts.csvREM.xml and news_reviews2009_2010_all.csvREM.xml are labeled by professionals of EBC and GCAO, respectively. Whereas, the other datasets are labeled by annotators who were employed by the authors. We strongly belief that it further requires enhancement.

Files

ebc_facebook_news_only_Amharic_texts.csvREM.xml

Files (9.2 MB)

Name Size
md5:4d74ce01f316b3a031038027ee33f7a9
1.3 MB Preview Download
md5:9aa49ebd3ee644e6c12c87359c19ff9c
683.6 kB Preview Download
md5:07b33febe06d5f9cb4509f31fb0143ad
6.8 MB Preview Download
md5:dbc0d1c3f7576ba9c1d744c75fc15ebf
473.3 kB Preview Download