Sentiment Analysis of 2024 Election Fraud Using SVM and Naïve Bayes Algorithms

Abstract

Elections are one of the main pillars of democracy, where the people's voice is the main determinant in government formation. Election fraud not only harms political competitors but also undermines public trust in democracy. The role of social media Twitter in widely disseminating information and disinformation adds to the challenge of maintaining election integrity. Sentiment analysis is the process of collecting and understanding individual opinions related to an event. Support Vector Machine (SVM) and Naïve Bayes algorithms are often used in this analysis due to their effectiveness and efficiency in text classification. This research aims to analyze public sentiment related to the 2024 presidential election fraud and compare the effectiveness of SVM and Naïve Bayes in sentiment classification. The study was conducted quantitatively, involving the stages of data collection, preprocessing, labeling, TF-IDF weighting, classification, and evaluation. The results of the sentiment analysis of public opinion on the 2024 presidential election fraud showed 42.5% negative sentiment, 38.6% neutral, and 18.9% positive. The dominance of negative sentiments reflects the public's concerns about election integrity. The high neutral sentiment indicates public doubt. To overcome this, transparency, strengthening supervisory institutions, electronic election technology, and strict law enforcement are needed. The SVM algorithm with RBF kernel produces 58% accuracy, better than Naïve Bayes with 51%.