Sentiment Analysis of Public Discourse on Education in Indonesia Using Support Vector Machine (SVM) and Natural Language Processing

Abstract

The growing influence of social media platforms like Twitter has transformed the landscape of public discourse, particularly on critical societal issues such as education. This study investigates public sentiment on education in Indonesia by analyzing tweets collected between January and June 2024, using sentiment analysis techniques powered by the Support Vector Machine (SVM) algorithm. By leveraging a dataset of 484 tweets, the analysis classified sentiments into positive, neutral, and negative categories, uncovering dominant patterns and their correlation with significant events in the education sector. The findings revealed a strong prevalence of neutral sentiments, emphasizing Twitter’s role as an informational hub. Positive sentiments were linked to public approval of equity-focused reforms, such as increased funding for rural schools, while negative sentiments reflected dissatisfaction with contentious policies, particularly standardized testing reforms. The study also examined temporal trends, identifying spikes in sentiment coinciding with major policy announcements, such as the significant surge in neutral sentiments in May 2024 following the government’s testing policy changes. These patterns illustrate the dynamic nature of public engagement on Twitter, shaped by real-time events and discussions. A comparative analysis with existing literature confirmed the value of social media as a barometer for public opinion, while also highlighting the unique context of Indonesian education. This research contributes to the growing field of digital society by demonstrating how sentiment analysis can provide actionable insights for policymakers and educators. It underscores the transformative potential of social media analytics in fostering inclusive and responsive governance. However, limitations such as dataset biases and classification challenges suggest avenues for future research, including multi-platform analysis and advanced natural language processing techniques. These findings serve as a foundation for leveraging sentiment analysis to enhance educational strategies and public communication in Indonesia's increasingly digitalized landscape.