Analisis Sentimen Ulasan Kepuasan Pengguna Aplikasi BSI Mobile Dengan Menggunakan Naive Bayes

Abstract

The BSI mobile application is a digital platform provided by BSI bank for carrying out financial transactions via smartphone. The BSI application does not only carry out transactions, however, it provides sharia features that are adapted to Islamic sharia principles such as sharia investment, zakat and infaq and so on. The methodological process consists of collecting data from 998 reviews which are used as a dataset, preprocessing with the aim of cleaning the data using several stages. Labeling is done by analyzing the sentiment of reviews on the dataset by dividing the data 60% to 40%, resulting in a positive value of 98% and a negative value of 2%. The implementation of the Naive Bayes algorithm in this research resulted in a confusion matrix evaluation with accuracy of 98.71%, precision of 99.22% and recall of 99.48%. This research aims to improve applications through reviews on Google Playstore by analyzing sentiment so that the service and quality of their applications match user feedback.