Sentiment Analysis of Social Media Platform Reviews Using the Naïve Bayes Classifier Algorithm

Abstract

The Covid-19 pandemic has caused significant changes in people's lifestyles which are further strengthened by the rapid development of technology. This has resulted in increased use of the internet and accelerated dissemination of information through social media platforms. Not only for self-expression, social media can also be a means of communication, information, education, and even used as a marketing tool. Several social media platforms have recently been popular and widely used, the number of users is increasing from year to year, and each user can provide a rating review of the application. To find out public opinion on social media platforms, sentiment analysis will be carried out on several social media platform applications on the Google Play Store, namely Twitter, Instagram and Tiktok which will later be used as material for evaluating these applications. In this study, the dataset was taken based on ratings from user reviews on the Google Play Store using the NBC (Naïve Bayes Classifier) method with the Python programming language. Based on testing of 1000 comment review data from each application, it was found that the majority gave positive sentiment (Twitter 57.2%, Instagram 74.1%, Tiktok 83.9%), and negative sentiment (Twitter 42.8%, Instagram 25.9%, Tiktok 16.1%) with an accuracy rate of 85.6% for the Twitter application, 83.6% for the Instagram application, and 84.8% for the Tiktok application.