Public Sentiment Analysis on the Inauguration of President Prabowo Subianto on Platform X Using the Support Vector Machine (SVM) Algorithm

Abstract

The inauguration of President Prabowo Subianto emerged as a pivotal political event that captured significant public interest and sparked a wide array of reactions across social media, particularly on the X platform (formerly known as Twitter). This research aims to categorize and analyze public sentiment regarding this historic moment by utilizing the Support Vector Machine (SVM) algorithm, a robust machine learning approach for classification tasks. A dataset comprising 1,000 tweets was initially gathered through targeted searches related to the inauguration. Subsequently, the data underwent a rigorous preprocessing phase, which included tokenization to break down text into individual components, stopword removal to eliminate irrelevant terms, filtering to exclude special characters and noise, and Term Frequency-Inverse Document Frequency (TF-IDF) transformation to convert textual data into a numerical format suitable for algorithmic processing. After preprocessing, 909 data points were prepared for further analysis. The dataset was then divided into two subsets: 80% allocated for training the model (727 data points) and 20% reserved for testing its performance (182 data points). The results of sentiment classification indicated that, among the test data, 653 tweets conveyed a positive sentiment toward the inauguration, whereas 74 tweets expressed a negative sentiment. Performance evaluation of the model demonstrated a commendable accuracy rate of 89.82%, alongside a precision of 89.82%, a recall of 100%, and an F1-score of 94.63%. These metrics highlight the model’s strong capability to accurately discern and classify public opinions related to political developments. The elevated recall rate, in particular, signifies the model’s ability to identify all instances of positive sentiment without omission. However, the precision score suggests some room for refinement in reducing misclassifications. The findings underscore the effectiveness of the SVM algorithm in dissecting and interpreting consumer sentiment toward significant political events. This provides a reliable tool for such analyses. Moreover, the outcomes of this study are anticipated to offer a valuable reference point for stakeholders and policymakers in leveraging data-driven approaches to gauge public opinion and monitor economic trends in Indonesia. This research also lays the groundwork for future investigations into sentiment analysis within the digital sphere. This could guide strategic communications and policy formulation based on real-time societal feedback