Sentiment Analysis of User Comments in the Threads Application on Google Play Store Using Multinomial Naive Bayes

Abstract

The advancement of technology, especially internet and social media technology, has enabled people across the country to connect and interact with each other. In this context, Instagram has become a popular platform for content creators to share their work. In an effort to compete with other platforms, Instagram launched an integrated app called Threads, which shares some features similar to Twitter. Threads allows users to share text-based posts and provides various other features. To enhance the quality of this application, developers need to review user comments. However, the influx of comments is substantial, making manual review inefficient. Therefore, an automated application is required to categorize comments and analyze user sentiment. By utilizing text mining techniques for sentiment analysis, developers can easily sort comments into positive and negative categories. Multinomial Naive Bayes was chosen as it's specifically designed for data with frequency occurrences, such as in text analysis. It is expected that this application can assist developers in improving the quality of the generated app. From the results of this study, an accuracy of 76% was achieved, which is  relatively good and offers potential for further development to attain better results.