Published May 12, 2020 | Version v1.0
Dataset Open

JULIELab/MEmoLon (Data)

  • 1. JULIE Lab, FSU Jena

Description

Overview

This deposit contains the resulting lexicons from our ACL 2020 paper "Learning and Evaluating Emotion Lexicons for 91 Languages". The main repository for this project – including models, experimental code, and analyses – can be found on GitHub or the associated zenodo deposit.

 

Content

This deposit includes four zip files, each one representing different versions of the lexicons. The one which we mainly refer to in the paper is MTL_grouped.zip. The other versions were employed as a baseline comparison (`ridge.zip`) or in a development experiment (all but `ridge.zip`). Each zip file contains 91 tsv files which are named <iso language code>.tsv. Please refer to the file lexicons_overview.csv to find the right code for your language.

Each tsv file constitutes a large-scale emotion lexicon for a particular language, covering roughly between 100k and 2M word type entries. Each word is described in terms of eight emotional variables: valence, arousal, dominance, joy anger, sadness, fear, and disgust.  

Please refer to our paper (see below) for additional details.

 

Citation

If you use our emotion lexicons in your work, please cite our paper: (coming up soon).

 

 

Files

lexicons_overview.csv

Files (9.5 GB)

Name Size Download all
md5:c206addf6b3015f53e28c22f6516e9de
1.8 kB Preview Download
md5:ca667a3adaf6edc8c6acabf04a1b86c7
2.4 GB Preview Download
md5:e773cbd8a8f3fc4b76a36c4b6d03fcd4
2.4 GB Preview Download
md5:499b8e48158ff4499fd9a41756e8b859
2.3 GB Preview Download
md5:32a8a1773671204f6fd761fe1439b69d
2.4 GB Preview Download