Sentiment Analysis Of State Officials News On Online Media Based On Public Opinion Using Naive Bayes Classifier Algorithm And Particle Swarm Optimization

Online media in general, ie all types of media formats that can only be accessed via the Internet contains text, photos, video, and sound. With the understanding of online media in general then email, mailing list (mailing list), website, blog, WhatsApp, and social media into the category of online media. the advantages of online media are compared to the conventional media due to its very flexible because it can be accessed in various places connected to the Internet network, the information provided is up to date and also anyone can comment on related articles. Information is spreading very quickly and is accompanied by freedom of expression resulting in different types of opinions, whether negative or positive. In this study, the authors apply the text mining process using Naive Bayes Classifier algorithm and Particle Swarm optimization to classify sentiments automatically. The data used are 200 comments online news portal readers about the sentiment to the attitude and actions of state officials in carrying out their duties. The data are classified manually and divided into each of 100 data for positive and negative sentiments. Then 200 data is used for testing. The results of this study resulted in a system that can classify sentiments automatically with accuracy ranging from 76.00%. With a fairly high degree of accuracy, Naive Bayes model supported by Particle Swarm optimization can provide solutions in classifying public opinion to the news of state officials to be more accurate and optimal.