Prediksi Elektabilitas Kandidat Presiden 2024 Menggunakan Algoritma Nave Bayes Classifier Pada Sosial Media Twitter

Sentiment analysis is a widely used tool among stakeholders to evaluate opinions on a subject. In this study, the object of investigation is sentiment analysis on a political figure, a candidate for the 2024 presidential election, which is currently a topic of discussion among netizens, especially on the Twitter platform. The issues raised include evaluating the performance of algorithms in sentiment classification, where some algorithms often exhibit low accuracy levels.
This research aims to understand public responses towards the 2024 election using the Twitter platform for sentiment analysis. The Naïve Bayes Classifier algorithm is employed to classify tweets into positive, negative, or neutral sentiments. A total of 1,444 data points were collected within the timeframe of December 15, 2023.
The test results indicate that the Naïve Bayes Classifier algorithm achieves a relatively high performance, with an average accuracy rate of 73% for Anies' data, 79.4% for Ganjar, and 79.8% for Prabowo.
This research concludes that public sentiment towards Ganjar is superior compared to others based on the Naïve Bayes Classifier method.
Keywords : Sentiment Analysis, Textblob, Naïve Bayes, Tweet Harvest, Presiden, Twitter.