Understanding Sentiment Across Genders: Challenges and Solutions

Hossein Hassani, Pouria Parvizi, Mohammad Yeganegi and Rangan Gupta
Additional contact information
Hossein Hassani: International Institute for Applied Systems Analysis (IIASA), Austria
Pouria Parvizi: Department of Computer Engineering, University of Kurdistan, Iran
Mohammad Yeganegi: International Institute for Applied Systems Analysis (IIASA), Austria

No 202515, Working Papers from University of Pretoria, Department of Economics

Abstract: The emergence of natural language processing models has made them a crucial part of financial and economic analysis, especially when it comes to understanding human behavior. Sentiment analysis has been used to understand people's opinions about policies, products, and services, as well as to gauge market sentiment. These insights can be used as input to other economic and financial models to enrich their performance. However, intertwining sentiment analysis with other analyses means that any inaccuracies or biases can negatively impact the final results, resulting in misleading conclusions. Specifically, any gender bias in sentiment analysis can lead to gender bias in subsequent analyses and final results. In this regard it is important to understand the potential sources of gender bias in sentiment analysis and address those biases. This study aims to provide a comprehensive understanding of gender bias in sentiment analysis and explore strategies to mitigate it, ensuring unbiased results.

Pages: 12 pages
Date: 2025-04
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:pre:wpaper:202515

Access Statistics for this paper

More papers in Working Papers from University of Pretoria, Department of Economics Contact information at EDIRC.
Bibliographic data for series maintained by Rangan Gupta ().