Analisis Sentimen Media Sosial terhadap Kebijakan Publik Menggunakan Metode Deep Learning

Abstract

The rapid development of information and communication technology has significantly increased the use of social media as a primary platform for expressing public opinions, criticisms, and support on various policy issues. Social media not only functions as a communication tool but also serves as a valuable data source for understanding public perception of government policies. Sentiment analysis has emerged as an effective approach to identify public opinion trends based on textual data generated by users. This study aims to analyze public sentiment toward a specific public policy using a deep learning approach to obtain a more accurate understanding of societal responses. Data were collected from social media platforms within a defined period. The research process included data collection, data cleaning, text normalization, tokenization, stopwords removal, and sentiment labeling into positive, negative, and neutral categories. The processed data were analyzed using a Long Short-Term Memory (LSTM) model due to its ability to capture contextual relationships between words. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results indicate that the LSTM model achieved high accuracy in sentiment classification. Negative sentiment was dominant in economic and public service issues, while positive sentiment appeared in policies with direct benefits. Social media data proved useful for real-time policy evaluation.