Predicting Sentiment in Yorùbá Written Texts: A Comparison of Machine Learning Models
| dc.contributor.author | Omolayo Abegunde | |
| dc.date.accessioned | 2025-05-27T16:25:18Z | |
| dc.date.available | 2025-05-27T16:25:18Z | |
| dc.date.issued | 2020 | |
| dc.description | The study presents a system that detects sentiment from Yorùbá sentences at the sentence level with a view to extracting the users’ opinion from the sentence. The study has pointed out various ways in which the work can be achieved using machine learning with different algorithms. The specific challenges encountered during the implementation of the SA for the Yorùbá language have been pointed out. The result shows that users’ opinions in Yorùbá sentences can be mined and the stand on an event can be determined. This study provides a holistic assessment of the SA system in Yorùbá; however, some sentences that express sarcasm were not captured in the study. The research concludes that NB outperforms other algorithms for the analysis of sentiments for Yorùbá sentences. Our future goal is to explore sentiment analysis in the area of named entity recognition and sarcasm for the Yorùbá language. | |
| dc.description.abstract | Sentiment analysis (SA) provides a rich set of tools and techniques for extracting and evaluating subjective information from large datasets. Users' opinions concerning an event are what determine the user perspective of such event, whether it is good or bad. This study compared three machine learning models (logistic regression, Naı̈ve Bayes, and support vector machine) with a view to identifying the best model for predicting sentiment in Yorùbá written texts at the sentence level. The corpus of Yorùbá records was created from several online and offline sources such as dictionaries, experts, the Bible, and social media, as well as the Awayoruba blog, and processed using Tákàdá. The system was implemented using the Python programming language and evaluated using mean opinion score and receiver operating characteristics. The research concludes that Naı̈ve Bayes (NB) outperforms other algorithm for analysis of sentiments for Yorùbá sentences. | |
| dc.description.sponsorship | Self | |
| dc.identifier.uri | 10.1007/978-3-030-29513-4_30 | |
| dc.identifier.uri | https://repository.run.edu.ng/handle/123456789/4822 | |
| dc.language.iso | en | |
| dc.publisher | Springer Nature Switzerland | |
| dc.relation.ispartofseries | 2 | |
| dc.title | Predicting Sentiment in Yorùbá Written Texts: A Comparison of Machine Learning Models | |
| dc.type | Article |