Citation

Abdul Rahman, Siti Soraya and Rusli, Amira An-Nur and Aga Mohd Jaladin, Rafidah (2025) Linguistic Indicators of Narcissistic Tendencies for Predicting Narcissistic Traits in Malaysians Through Social Media Content. Journal of Communication, Language and Culture, 5 (2). pp. 98-112. ISSN 2805-444X

Text 14876.pdf - Published Version Restricted to Repository staff only Download (700kB)

Abstract

Narcissistic tendencies are increasingly observable in online interactions, particularly on social media platforms. X (formerly Twitter) has been identified as a popular platform among individuals exhibiting narcissistic traits. Predicting these traits based on language use remains an essential area of study. This research examines narcissistic tendencies among Malaysian social media users by analysing 2,129 posts on Twitter. Utilising natural language processing (NLP) methods and machine learning algorithms, we assess linguistic patterns, sentiment, and engagement metrics to identify indicators of narcissism. Four machine learning algorithms, Support Vector Machine (SVM), Naïve Bayes, Logistic Regression, and Gradient Boosting, were assessed based on multiple performance metrics. Results indicate that SVM is the most effective model, achieving 80% accuracy with 10-fold cross-validation, demonstrating its reliability in predicting narcissistic traits. These findings contribute to computational psychology and social media analytics, offering insights into the psychological dimensions of digital self-presentation and the implications of narcissistic behaviours in online communities.

Item Type:	Article
Uncontrolled Keywords:	Narcissism, social media, X (formerly Twitter), machine learning, natural language processing
Subjects:	B Philosophy. Psychology. Religion > BF Psychology (General) > BF1-990 Psychology Q Science > Q Science (General) > Q300-390 Cybernetics
Depositing User:	Nurin Syazwani Azmi
Date Deposited:	11 Nov 2025 02:23
Last Modified:	11 Nov 2025 02:23
URII:	http://shdl.mmu.edu.my/id/eprint/14876

Downloads

Edit (login required)