Analysis ANALYSIS OF SENTIMENT IN TWEETS OF POLRI PR DIVISION USING THE METHOD NAÏVE BAYES CLASSIFICATION

Abstract

News is a story or information about an event that is hotly discussed. It can be utilized by online media to provide updated news to the public. The use of social media among the public in disseminating information is fast so that distance and time are not an obstacle for social media users to be able to receive and access information developments widely without constraints. Twitter is a social media platform that is widely used by individuals, governments, and organizations, including the police, to publish news about related agencies. By applying one of the functions of text mining, namely the Naïve Bayes Classifier to analyze public sentiment towards tweets from the Indonesian Police Public Relations Division’s Twitter account, manually analyzing sentiment on tweets is no longer effective, so a Naïve Bayes Classifier method is needed to automatically analyze sentiment on tweets into positive and negative opinions. So, by using this algorithm, this study gets a fairly high accuracy value. In this study, a high accuracy value of 76% when splitting 10% of testing data and 90% of training data. But when preprocessing and    implementing new data in the streamlit framework, it takes up to 1 minute to process the data.