Overview of the EmoEvalEs task on emotion detection for Spanish at IberLEF 2021

Resumen

This paper presents the EmoEvalEs shared task, organized at IberLEF 2021, as part of the 37th International Conference of the Spanish Society for Natural Language Processing (SEPLN 2021). The aim of this task is to promote the Emotion detection and Evaluation for Spanish. It consists of a fine-grained emotion classification of tweets from the EmoEvalEs corpus in one of these seven classes: anger, disgust, fear, joy, sadness, surprise, or others. In this edition, 70 teams registered, 15 submitted results and 11 presented papers describing their systems. Most teams experimented with neural networks, being Transformers the most widely used model. It should be noted that few of them also considered the features of offensiveness and event that were provided in the corpus apart from the tweet texts.