Home // HUSO 2019, The Fifth International Conference on Human and Social Analytics // View article

Towards an Empirically-Grounded Framework for Emotion Analysis

Authors:
Luna De Bruyne
Orphée De Clercq
Véronique Hoste

Keywords: Emotion Detection, NLP, Emotion Annotation

Abstract:
The first step in training a system for automatic emotion detection consists of manual data annotation. Because there is no consensus on a standard emotion framework, we established a label set which is justified both theoretically and practically. Frequency and cluster analysis of 229 tweet annotations resulted in a label set containing the 5 emotions Love, Joy, Anger, Nervousness and Sadness. Our label set shows fair resemblance to Ekman's basic emotions, but due to our data-driven approach, our label set is much more grounded in the task (emotion detection) and the domain (Dutch tweets).

Pages: 11 to 16

Copyright: Copyright (c) IARIA, 2019

Publication date: June 30, 2019

Published in: conference

ISSN: 2519-8351

ISBN: 978-1-61208-725-2

Location: Rome, Italy

Dates: from June 30, 2019 to July 4, 2019