ANALISIS SENTIMEN UNTUK MEMPREDIKSI HASIL CALON PEMILU PRESIDEN MENGGUNAKAN LEXICON BASED DAN RANDOM FOREST

Abstract

The Presidential Election is one of the crucial moments in Indonesian politics. To predict the election results, sentiment analysis methods can be used to evaluate public opinions through social media. One of the popular social media platforms nowadays is Twitter. As the Republic of Indonesia's Presidential Election approaches, there is an increasing number of tweets discussing the event. This situation creates a favorable opportunity to conduct sentiment analysis on the election campaign topic. There are various opinions from Twitter users with positive, neutral, and negative sentiments. The collected tweet data undergoes preprocessing, involving two main processes: cleaning and stemming. Therefore, sentiment analysis is necessary to understand the public's tendencies towards the election. The objective of this research is to obtain sentiment analysis of the text documents to determine positive or negative sentiments. Two methods, namely Random Forest and Lexion Based, are used to predict the presidential candidates' sentiments. Random Forest is employed to analyze sentiments in text data collected from various sources, including social media, news websites, and online forums. This method involves an ensemble of decision trees working collectively to classify sentiments as positive, negative, or neutral towards the Presidential Election candidates. On the other hand, Lexion Based is used to associate words in the text with specific sentiments.