Analysis of Twitter Netizens' Sentiment Regarding the Allocation of Untargeted KIP Scholarships Using the Naive Bayes Classifier Algorithm

The existence of the Kartu Indonesia Pintar (KIP) or smart Indonesia Card program can help underprivileged people to continue their education in higher education. However, there are still many problems in the allocation of scholarship recipients. One of the issues that is currently being discussed on social media X or Twitter, is about the allocation of KIP scholarships that are not on target. The public can have various sentiments or responses to the allocation of the scholarship, which can be categorized into positive and negative sentiments. This study is an analysis of the sentiment of KIP allocation on Twitter social media to classify tweets and comments with negative and positive content using the Naïve Bayes classifier method. The research flow uses the SEMMA (Sample, Explore, Modify, Model, and Assess) method. The dataset used in this study is the collection of datasets through crawled tweets using the Python scrapper in google colab and then selecting important attributes at the exploration stage. The crawling data starts from December 12, 2023 to May 18, 2024. Sentiment is divided into positive and negative. The results of the interpretation show that the majority of tweets analyzed in this dataset have positive sentiment, with a comparison of 66.7% positive sentiment and 33.3% negative sentiment. From the results of the model test, the accuracy obtained was 0.75 for Gaussian Naïve Bayes, 0.67 for Multinomial Naïve Bayes, and 0.67 for Bernoulli naïve Bayes. Thus, it can be concluded that Gaussian Naïve Bayes is more effective in classifying tweet sentiment in the dataset tested, compared to Multinomial and Bernoulli Naïve Bayes.