Klasfikasi Sentimen Aplikasi X Terhadap Gugatan Pemilu 2024 Menggunakan Naïve Bayes dan Textblob

Abstract

This study analyzes public sentiment towards the 2024 Election Results Dispute at the Constitutional Court through the X (Twitter) application using the Naïve Bayes and TextBlob methods. The dataset was collected through crawling and preprocessing to remove duplicate data, clean, and normalize the tweets. Labeling was done using TextBlob, followed by sentiment classification using the Naïve Bayes algorithm. The results show that out of 898 tweets analyzed, the TextBlob labeling identified 340 positive tweets, 427 neutral tweets, and 131 negative tweets. Meanwhile, the Naïve Bayes classification resulted in 515 positive tweets, 281 neutral tweets, and 102 negative tweets, demonstrating high accuracy with 95.29%. Data visualization through word clouds and bar charts helped map the sentiment distribution clearly. These findings provide valuable insights into public opinion on the election results dispute, with the majority of sentiments being positive and neutral.