Selection of Dataset for Emotion Detection with Respect to Federated Learning

G. K. Jakir Hussain, G. Manoj

2025

Abstract

Emotion detection (ED) plays a vital role in applications like health, human–computer interaction (HCI), and personalization. Federated learning (FL) has the potential to provide robust models for ED while keeping data private across distributed clients. This work has considered the critical factors that influence dataset selection for FL on ED tasks. The range of dataset types, class balance for varying emotional states, and attention to privacy considerations that safeguard users' sensitive information are among the important factors to consider. Therefore, several datasets are examined in terms of how well they extent the range of emotional expressions and replicate actual client data distributions. This study highlights datasets such as FER-2013 for photos, RAVDESS for audio, and ISEAR for textual data are highly relevant for the construction of generalized ED models in FL setups that simulate near-real-world settings, as per review publications. The FL technique can balance the data diversity and privacy preservation and hence saves a pathway to harness collective intelligence from distributed data sources while ensuring ethical practices for handling data. Finally, FL is becoming a critical topic for addressing issues in identifying ED by selecting the suitable datasets.

Paper Citation

in Harvard Style

Hussain G. and Manoj G. (2025). Selection of Dataset for Emotion Detection with Respect to Federated Learning. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 593-599. DOI: 10.5220/0013917400004919

in Bibtex Style

@conference{icrdicct`2525,
author={G. Hussain and G. Manoj},
title={Selection of Dataset for Emotion Detection with Respect to Federated Learning},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={593-599},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013917400004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Selection of Dataset for Emotion Detection with Respect to Federated Learning
SN - 978-989-758-777-1
AU - Hussain G.
AU - Manoj G.
PY - 2025
SP - 593
EP - 599
DO - 10.5220/0013917400004919
PB - SciTePress