Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets

Social media serve as a crucial platform for expressing opinions and perspectives. Its texts often characterised by code-switching or mixed languages in multilingual setting. This results in a diverse and complex linguistic context, which can negatively affect the accuracy of sentiment analysis for...

Full description

Saved in:
Bibliographic Details
Main Authors: Yusuf Aliyu, Aliza Sarlan, Kamaluddeen Usman Danyaro, Abdullahi Sani abd Rahman, Aminu Aminu Muazu, Mustapha Yusuf Abubakar
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:International Journal of Information Management Data Insights
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2667096825000126
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849333713027989504
author Yusuf Aliyu
Aliza Sarlan
Kamaluddeen Usman Danyaro
Abdullahi Sani abd Rahman
Aminu Aminu Muazu
Mustapha Yusuf Abubakar
author_facet Yusuf Aliyu
Aliza Sarlan
Kamaluddeen Usman Danyaro
Abdullahi Sani abd Rahman
Aminu Aminu Muazu
Mustapha Yusuf Abubakar
author_sort Yusuf Aliyu
collection DOAJ
description Social media serve as a crucial platform for expressing opinions and perspectives. Its texts often characterised by code-switching or mixed languages in multilingual setting. This results in a diverse and complex linguistic context, which can negatively affect the accuracy of sentiment analysis for low-resource languages such as Hausa. Prior research has predominantly concentrated on sentiment analysis within single-language data rather than code-switched data. This paper proposes an efficient hyperparameter tuning framework and a novel stemming algorithm for the Hausa language. The framework leverages word embeddings to determine the polarity scores of code-mixed tweets and enhances the accuracy of sentiment analysis models in low-resource language. The extensive experiments demonstrate the framework's efficiency and reveal a superior performance of transformer models over conventional deep learning models. The framework achieves a balance between accuracy and computational efficiency, making it suitable for deployment in practical applications. Compared to state-of-the-art transformer models, our framework significantly reduces computational costs while maintaining competitive performance. Notably, the AfriBERTa model achieves outstanding results, with an F1-score of 0.92 and an accuracy of 0.919, surpassing current baseline standards. These findings have broad implications for social media monitoring, customer feedback analysis, and public sentiment tracking, enabling more inclusive and accessible NLP tools for underrepresented linguistic communities.
format Article
id doaj-art-b8a2193d76e349f7b3babadbe9e7ce5d
institution Kabale University
issn 2667-0968
language English
publishDate 2025-06-01
publisher Elsevier
record_format Article
series International Journal of Information Management Data Insights
spelling doaj-art-b8a2193d76e349f7b3babadbe9e7ce5d2025-08-20T03:45:47ZengElsevierInternational Journal of Information Management Data Insights2667-09682025-06-015110033010.1016/j.jjimei.2025.100330Deep learning techniques for sentiment analysis in code-switched Hausa-English tweetsYusuf Aliyu0Aliza Sarlan1Kamaluddeen Usman Danyaro2Abdullahi Sani abd Rahman3Aminu Aminu Muazu4Mustapha Yusuf Abubakar5Department of Computer and Information Science, Universiti Teknologi PETRONAS, Seri Iskandar 32610 Perak, Malaysia; Corresponding authors.Center for Foundation Studies, Universiti Teknologi PETRONAS, Seri Iskandar, Perak, 32610, MalaysiaDepartment of Computer and Information Science, Universiti Teknologi PETRONAS, Seri Iskandar 32610 Perak, Malaysia; Centre for Cyber-Physical Systems (C2PS), Universiti Teknologi PETRONAS, Seri Iskandar, Perak 32610, Malaysia; Corresponding authors.Department of Computer and Information Science, Universiti Teknologi PETRONAS, Seri Iskandar 32610 Perak, MalaysiaDepartment of Computer and Information Science, Universiti Teknologi PETRONAS, Seri Iskandar 32610 Perak, Malaysia; Computer Science Department, Faculty of Natural and Applied Science, Umaru Musa Yar'adua University, Katsina, NigeriaComputer Science Department, School of Technology, Kano State Polytechnic, Kano 700231, NigeriaSocial media serve as a crucial platform for expressing opinions and perspectives. Its texts often characterised by code-switching or mixed languages in multilingual setting. This results in a diverse and complex linguistic context, which can negatively affect the accuracy of sentiment analysis for low-resource languages such as Hausa. Prior research has predominantly concentrated on sentiment analysis within single-language data rather than code-switched data. This paper proposes an efficient hyperparameter tuning framework and a novel stemming algorithm for the Hausa language. The framework leverages word embeddings to determine the polarity scores of code-mixed tweets and enhances the accuracy of sentiment analysis models in low-resource language. The extensive experiments demonstrate the framework's efficiency and reveal a superior performance of transformer models over conventional deep learning models. The framework achieves a balance between accuracy and computational efficiency, making it suitable for deployment in practical applications. Compared to state-of-the-art transformer models, our framework significantly reduces computational costs while maintaining competitive performance. Notably, the AfriBERTa model achieves outstanding results, with an F1-score of 0.92 and an accuracy of 0.919, surpassing current baseline standards. These findings have broad implications for social media monitoring, customer feedback analysis, and public sentiment tracking, enabling more inclusive and accessible NLP tools for underrepresented linguistic communities.http://www.sciencedirect.com/science/article/pii/S2667096825000126Sentiment analysisLow-resourceCode-switchedHausa languageWord-embeddingTransformer
spellingShingle Yusuf Aliyu
Aliza Sarlan
Kamaluddeen Usman Danyaro
Abdullahi Sani abd Rahman
Aminu Aminu Muazu
Mustapha Yusuf Abubakar
Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets
International Journal of Information Management Data Insights
Sentiment analysis
Low-resource
Code-switched
Hausa language
Word-embedding
Transformer
title Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets
title_full Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets
title_fullStr Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets
title_full_unstemmed Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets
title_short Deep learning techniques for sentiment analysis in code-switched Hausa-English tweets
title_sort deep learning techniques for sentiment analysis in code switched hausa english tweets
topic Sentiment analysis
Low-resource
Code-switched
Hausa language
Word-embedding
Transformer
url http://www.sciencedirect.com/science/article/pii/S2667096825000126
work_keys_str_mv AT yusufaliyu deeplearningtechniquesforsentimentanalysisincodeswitchedhausaenglishtweets
AT alizasarlan deeplearningtechniquesforsentimentanalysisincodeswitchedhausaenglishtweets
AT kamaluddeenusmandanyaro deeplearningtechniquesforsentimentanalysisincodeswitchedhausaenglishtweets
AT abdullahisaniabdrahman deeplearningtechniquesforsentimentanalysisincodeswitchedhausaenglishtweets
AT aminuaminumuazu deeplearningtechniquesforsentimentanalysisincodeswitchedhausaenglishtweets
AT mustaphayusufabubakar deeplearningtechniquesforsentimentanalysisincodeswitchedhausaenglishtweets