A Comprehensive Review on Sentiment Analysis in Low-Resource and Code-Mixed Languages

View at publisher Summarise with VIRAT
Authors: Gautham Vadakkekara Suresh, Nikita
Cited by: 0
References: 17

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

As the world is getting digital day by day, sentiment analysis in data scarce and code-mixed languages has become crucial since more people are expressing their opinions on various digital platforms, along with mixing English with their local languages. Since there are very limited dataset, repeated transliteration on these regional languages, producing a sentiment classification would be more difficult compared to resource-rich languages. Because of these reasons, currently existing models face problems to maintain robust performance, especially in social media, feedback of customers where code-mixing is usual. This paper therefore reviews 20 papers on sentiment analysis for low-resource and code-mixed languages, focusing on different datasets used, various model building approaches and their evaluation strategies. A growing trend towards neural and transformer-based architectures, as well as problems occurred due to low training data availability in some languages, insufficient labelled corpora, differences of sentiment across different contexts were discovered after study of these papers. The primary objective of this review is to consolidate the current findings to address methodological gaps while providing future directions to advance sentiment analysis across neglected linguistic contexts.

Fields of Study & Subjects

Sentiment Analysis and Opinion Mining Topic Modeling Hate Speech and Cyberbullying Detection

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
915–922
Publication date
21 January 2026
Language
EN