Abstract
The fast rise of social media platforms has given academics unparalleled access to user-generated data, allowing for large-scale studies of public attitudes and moods. In a nation with such a rich culture as Malaysia, it is typical to see tweets written in Malay, local slang, and English. This language variation makes it more difficult to analyze emotions, especially given the need for labeled data required for supervised learning approaches. This study examines and categorizes Malaysians' mood expressions, mostly using code-mixing techniques discovered on Twitter. The study uses the Jupyter Notebook application to visualize and analyze a dataset comprising 2184 out of 2190 Twitter tweets after data pre-processing. The NRCLex Affect lexicon is used for both data analysis and emotion classification. The analysis reveals that approximately 50.9% of Twitter users were likely to express happiness, followed by 19.3% expressing trust, 10.9% expressing fear, 13.1% expressing sadness, 3.1% expressing anger, and 2.7% expressing surprise. The results are promising, as a relatively high level of accuracy was achieved even with a small initial labeled dataset. This outcome is significant when labeled datasets for emotion analysis are limited. Additionally, the research provides real-time analysis of emotions. The successful classification of mood expression in code-mixed tweets provides insights into Malaysians' emotional states, contributing to a deeper understanding of public sentiment. Understanding the prevailing mood is valuable in gauging public opinion, assessing social trends, and informing decision-making processes at both individual and societal levels.
Author Information
Latifah Abd Latib, Universiti Selangor, Malaysia
Hema Subramaniam, Universiti Malaya, Malaysia
Affezah Ali, Taylor's University, Malaysia
Siti Khadijah Ramli, Universiti Selangor, Malaysia
Paper Information
Conference: BAMC2023
Stream: Communication
This paper is part of the BAMC2023 Conference Proceedings (View)
Full Paper
View / Download the full paper in a new tab/window
To cite this article:
Latib L., Subramaniam H., Ali A., & Ramli S. (2023) Unveiling Mood Classifications in Malaysia: Analysing Code-Mixed Twitter Data for Emotional Expression ISSN: 2435-9475 – The Barcelona Conference on Arts, Media & Culture 2023: Official Conference Proceedings https://doi.org/10.22492/issn.2435-9475.2023.19
To link to this article: https://doi.org/10.22492/issn.2435-9475.2023.19
Comments
Powered by WP LinkPress