Analisis Sentimen Ulasan Pengguna Pada Aplikasi Wondr By BNI Menggunakan Metode Klasifikasi Algoritma Naïve Bayes

Abstract

To better serve its customers and make financial transactions easier, BNI Bank developed Wondr by BNI, a type of digital innovation in the banking industry. Users may access and download the Wondr by BNI application via the App Store for iOS devices and the Google Play Store for Android devices. To gauge the app's usefulness, reviews are required, and reviews can greatly impact the app's future updates. Consequently, the purpose of this research is to ascertain, by sentiment analysis, if reviews of the Wonder by BNI application are favorable or negative. The data used for this study came from the Google Play Store, and it was labeled using Microsoft Excel to get the positive and negative review counts. Then, it was processed with Data Preprocessing, and then weighting was done with TF-IDF to get the word counts. The data was then classified using the Naïve Bayes algorithm, and finally, it was evaluated to find out how accurate the results were. The dates of this study's execution are September 23, 2024, and January 10, 2025. Based on the data, Naïve Bayes produced rather accurate predictions, with a recall of 86.55%, a precision of 80.85%, and an accuracy rating of 83.03%. Because of its strong recall, the model was successful in identifying a small number of genuine positive cases.