<![CDATA[With the increasing use of mobile banking applications in Indonesia, understanding user reviews and feedback has become increasingly important for banks to enhance the services and performance of the applications they offer. This research aims to analyze user sentiment towards the mobile banking applications BCA, BNI, Brimo, and Byond by BSI, and to compare the effectiveness of the Random Forest, K-Nearest Neighbor (KNN), Naive Bayes, and Support Vector Machine (SVM) algorithms. The data used consists of user reviews obtained from Google Play Store through web scraping techniques, with 4,000 samples of reviews divided into training data (80%) and testing data (20%). The pre-processing process is conducted to prepare the data, which includes stopword removal and tokenization, using the Bag of Words (BoW) method.Based on the labeling results that can be seen in the visualization stage, it is known that the Byond by BSI Mobile application has a positive sentiment with 540 more reviews and a negative sentiment with 528 fewer reviews compared to other mobile banking applications. In the form of a comparative matrix graph, the Random Forest algorithm has a higher accuracy value of 0.58 for the BCA application and 0.74 for the Brimo application, while Naive Bayes has an accuracy value of 0.71, which is greater for the BNI mobile banking application, and Support Vector Machine has an accuracy value of 0.74, which is higher for the Byond by BSI mobile banking application. From the explanations above, it means that the Random Forest algorithm is capable of classifying efficiently and effectively compared to the other three algorithms. With the results of this research, it is hoped to provide important insights for mobile banking application developers to improve service quality based on user feedback, as well as to recommend the use of Random Forest for more accurate and reliable sentiment analysis.]]>
Copyrights © 2025
