A MULTIMODAL SENTIMENT ANALYSIS OF INTERNET MEMES IN THE COURSE OF PHILIPPINE ELECTIONS

As internet memes increasingly dominate political communication, concerns about their role in reinforcing biases, fostering polarization, and oversimplifying complex political issues have intensified. Employing Natural Language Processing (NLP) in the context of Sentiment Analysis, the Darth Sentiment Analyzer—a dedicated web-based software based on Valence Aware Dictionary and sEntiment Reasoner (VADER) —was utilized to examine the sentiments (positive, negative, neutral) expressed in a dataset of over 20,000 memes, revealing that neutral sentiment is the most prevalent, followed by positive and negative tones. The study applies Halliday’s Systemic Functional Linguistics and Kress and Van Leeuwen’s Grammar of Visual Design to analyze both the textual and visual elements of memes, providing deeper insights into how these components collaboratively influence public sentiment. Key findings demonstrate significant variation in sentiment across political affiliations, with liberalism being the most prominent ideology, while fascism and green ideologies are underrepresented. The study also identifies the ideological appeals that shape political engagement in online spaces, with implications for educational practices. The research contributes to the growing body of research on the intersection of technology, politics, and education, emphasizing the need for responsible digital citizenship in an era of widespread online influence. KEYWORDS: Education, Applied linguistics, Internet memes, natural language processing, sentiment analysis, Filipino voters, Philippines