English

Dialog speech sentiment classification for imbalanced datasets

Computation and Language 2021-09-16 v1 Artificial Intelligence

Abstract

Speech is the most common way humans express their feelings, and sentiment analysis is the use of tools such as natural language processing and computational algorithms to identify the polarity of these feelings. Even though this field has seen tremendous advancements in the last two decades, the task of effectively detecting under represented sentiments in different kinds of datasets is still a challenging task. In this paper, we use single and bi-modal analysis of short dialog utterances and gain insights on the main factors that aid in sentiment detection, particularly in the underrepresented classes, in datasets with and without inherent sentiment component. Furthermore, we propose an architecture which uses a learning rate scheduler and different monitoring criteria and provides state-of-the-art results for the SWITCHBOARD imbalanced sentiment dataset.

Keywords

Cite

@article{arxiv.2109.07228,
  title  = {Dialog speech sentiment classification for imbalanced datasets},
  author = {Sergis Nicolaou and Lambros Mavrides and Georgina Tryfou and Kyriakos Tolias and Konstantinos Panousis and Sotirios Chatzis and Sergios Theodoridis},
  journal= {arXiv preprint arXiv:2109.07228},
  year   = {2021}
}

Comments

To be published in SPECOM & ICR 2021 Electronic Proceedings by the Springer Nature

R2 v1 2026-06-24T05:59:05.612Z