Analisis Sentimen Terhadap Isu Kecurangan Pemilu 2024 Pada Platfom Twitter (X) Dengan Metode Naive Bayes Multinomial Dan Cosine Similiarity

Abstract

In an increasingly complex digital era, sentiment analysis has become a vital instrument in understanding the nuances of public opinion. This technique, which utilizes artificial intelligence and Machine Learning, allows us to extract knowledge about people's attitudes, emotions and perceptions towards various issues. This research examines public sentiment regarding the issue of fraud in the 2024 Election on the social media platform Twitter using a text mining-based sentiment analysis approach. Data was obtained through a crawling process using the Python programming language. The research methodology includes a series of stages, starting from data cleaning to improve quality, continuing with word weighting using the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm, and ending with modeling using the Naïve Bayes Classifier algorithm. Model evaluation was carried out systematically by applying the Naive Bayes, Confusion Matrix and K-Fold Cross Validation methods to measure the level of accuracy and effectiveness of the model developed. This research aims to produce in-depth knowledge regarding the trends and dynamics of public sentiment regarding the issue of fraud in the 2024 Election in the realm of social media, especially Twitter (X). Based on the research results, it shows a percentage of 67.7%.