Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Emotion Analysis in NLP: Trends, Gaps and Roadmap for Future Directions

Flor Miriam Plaza-del-Arco, Alba Curry, Amanda Cercas Curry, Dirk Hovy

TL;DR

This paper surveys 154 NLP publications (2014–2022) to examine how emotion analysis (EA) tasks are defined, which emotion frameworks are used, how subjective and cultural factors are considered, and which applications are pursued. It finds a heavy reliance on Ekman-based discrete emotions, limited attention to demographic and cultural variation, inconsistent terminology, and a lack of interdisciplinarity, which together hinder cross-task comparability and generalization. A four-part roadmap is proposed: incorporate demographic diversity, tailor emotion categories to the task, standardize EA nomenclature, and foster interdisciplinarity, in order to enable more nuanced, domain-appropriate, and culturally aware EA in NLP. The work aims to guide dataset creation, model design, and evaluation toward more inclusive and context-sensitive emotion modeling in NLP.

Abstract

Emotions are a central aspect of communication. Consequently, emotion analysis (EA) is a rapidly growing field in natural language processing (NLP). However, there is no consensus on scope, direction, or methods. In this paper, we conduct a thorough review of 154 relevant NLP publications from the last decade. Based on this review, we address four different questions: (1) How are EA tasks defined in NLP? (2) What are the most prominent emotion frameworks and which emotions are modeled? (3) Is the subjectivity of emotions considered in terms of demographics and cultural factors? and (4) What are the primary NLP applications for EA? We take stock of trends in EA and tasks, emotion frameworks used, existing datasets, methods, and applications. We then discuss four lacunae: (1) the absence of demographic and cultural aspects does not account for the variation in how emotions are perceived, but instead assumes they are universally experienced in the same manner; (2) the poor fit of emotion categories from the two main emotion theories to the task; (3) the lack of standardized EA terminology hinders gap identification, comparison, and future goals; and (4) the absence of interdisciplinary research isolates EA from insights in other fields. Our work will enable more focused research into EA and a more holistic approach to modeling emotions in NLP.

Emotion Analysis in NLP: Trends, Gaps and Roadmap for Future Directions

TL;DR

This paper surveys 154 NLP publications (2014–2022) to examine how emotion analysis (EA) tasks are defined, which emotion frameworks are used, how subjective and cultural factors are considered, and which applications are pursued. It finds a heavy reliance on Ekman-based discrete emotions, limited attention to demographic and cultural variation, inconsistent terminology, and a lack of interdisciplinarity, which together hinder cross-task comparability and generalization. A four-part roadmap is proposed: incorporate demographic diversity, tailor emotion categories to the task, standardize EA nomenclature, and foster interdisciplinarity, in order to enable more nuanced, domain-appropriate, and culturally aware EA in NLP. The work aims to guide dataset creation, model design, and evaluation toward more inclusive and context-sensitive emotion modeling in NLP.

Abstract

Emotions are a central aspect of communication. Consequently, emotion analysis (EA) is a rapidly growing field in natural language processing (NLP). However, there is no consensus on scope, direction, or methods. In this paper, we conduct a thorough review of 154 relevant NLP publications from the last decade. Based on this review, we address four different questions: (1) How are EA tasks defined in NLP? (2) What are the most prominent emotion frameworks and which emotions are modeled? (3) Is the subjectivity of emotions considered in terms of demographics and cultural factors? and (4) What are the primary NLP applications for EA? We take stock of trends in EA and tasks, emotion frameworks used, existing datasets, methods, and applications. We then discuss four lacunae: (1) the absence of demographic and cultural aspects does not account for the variation in how emotions are perceived, but instead assumes they are universally experienced in the same manner; (2) the poor fit of emotion categories from the two main emotion theories to the task; (3) the lack of standardized EA terminology hinders gap identification, comparison, and future goals; and (4) the absence of interdisciplinary research isolates EA from insights in other fields. Our work will enable more focused research into EA and a more holistic approach to modeling emotions in NLP.
Paper Structure (22 sections, 4 figures, 1 table)

This paper contains 22 sections, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Survey of Emotion Analysis in NLP
  4. Nomenclature
  5. Emotion models and frameworks
  6. Data
  7. Applications
  8. Shortcomings of EA in NLP
  9. Lack of Diversity in Available Datasets
  10. Diversity in emotion labels.
  11. Diversity in annotation schemes.
  12. Inconsistency in Terminology
  13. Lack of Interdisciplinarity
  14. Demographics and Cultural Implications
  15. Discussion
  16. ...and 7 more sections

Figures (4)

  • Figure 1: Variation in emotion annotation based on demographics. Annotators with distinct demographic profiles, a 60-year-old male and a 20-year-old male, exhibit varying interpretations of the emotions evoked.
  • Figure 2: Distribution of papers considered in our survey by year and keyword.
  • Figure 3: Most common nomenclature pairs for EA tasks used in the literature.
  • Figure 4: Most common emotion categories used in the EA literature.