Sentiment Analysis by Using Naïve Bayes Classification and Support Vector Machine, Study Case Sea Bank

Abstract

Information technology is developing at a rapid pace, changing people's lives, particularly in the financial sector where customer demands are rising, and banks must innovate to convert from traditional to technological banking systems while also increasing competency and efficiency through improved services. Innovations in digital banking have arisen in Indonesia as a result of technical progress. SEA Bank is one such digital bank; it was established in Indonesia in 2021. An app that may be found on the Google Play Store is used for all transactions. However, there are instances when the application's performance falls short of users' expectations, which prompts some users to voice their dissatisfaction. In order to determine if the evaluations are either beneficial or detrimental, the author therefore carried out a sentiment analysis study on SEA Bank using the Naïve Bayes classification and Support Vector Machine techniques. This was then implemented on a website utilizing the Flask framework. In the experiments with 90% training data, 10% testing data, and k = 10, the results of this study demonstrated that the sentiment classification process using the SVM algorithm was the best classification algorithm for evaluating its accuracy, precision, and recall values of 93.99%, 94.60%, 98.87%, and an F1 score of 96.69%.