Analysis of Public Sentiment Towards President Prabowo's Work Program Using The CNN

Abstract

Digital media has now become the primary means for Indonesians to receive and respond to information, including the work programs presented by Prabowo Subianto. One of the programs that is widely discussed by the public is related to efforts to improve the national economy. Public responses to this issue are widespread on social media, reflecting diverse sentiments. Therefore, this study aims to analyze the sentiment of comments from social media users X regarding President Prabowo's work programs in the economic sector, using a deep learning approach based on the Convolutional Neural Network (CNN) architecture. The methods employed include data collection, text preprocessing, and training a CNN model. The dataset used consisted of 2,467 data points, with 1,086 labeled as positive and 1,381 labeled as negative. The test results showed that the model achieved an accuracy of 87.45% and an Area Under the Curve (AUC) score of 0.9373, indicating excellent classification performance in distinguishing between positive and negative sentiments. This study proves that the combination of CNN and FastText is a practical approach to understanding text-based public opinion from social media.