ANALYSIS OF PUBLIC SENTIMENT TOWARDS 2024 PRESIDENTIAL CANDIDACY USING NAÏVE BAYES ALGORITHM

Abstract

This study analyzes public sentiment towards presidential nominations using text mining techniques and machine learning. The dataset consists of 670 tweets collected from social media. The analysis process includes a data pre-processing phase, encompassing text cleaning, case folding, tokenization, stopword removal, and stemming using the Sastrawi library for the Indonesian language. Sentiment labeling was was performed using NLTK's SentimentIntensityAnalyzer, categorizing tweets into positive, negative, or neutral sentiments. The analysis results reveal the sentiment distribution among the analyzed tweets. Data modeling was performed using the Naive Bayes algorithm, which achieved an accuracy of 97.78% on the Iris dataset as an implementation example. The confusion matrix and classification report demonstrate the model's excellent performance in distinguishing sentiment classes. This research provides insights into public opinion regarding presidential nominations and demonstrates the effectiveness of text mining techniques and machine learning in sentiment analysis. The method can be applied to understand public opinion trends in other political and social contexts