Analisis Sentimen Ulasan Pengguna Aplikasi Livin’ by Mandiri dengan Menggunakan Algoritma Support Vector Machine dan Naïve Bayes

Abstract

Application Living in Mandiri is one of the applications of mobile banking popular in Indonesia. Based on Mark Top Brand Index (TBI), Livin' by Mandiri is in the third category of mobile banking Top Brand Award 2023. Even though it is popular, the application accepts Lots of reviews that can influence the image and satisfaction among Customers. Research This analyzes sentiment review users on Google Play using the algorithm Support Vector Machine (SVM) and Naïve Bayes and gives repair quality to the application. Primary data in the study This originates from observation and web scraping, while secondary data originates from study literature. Analysis process sentiment was done with preprocessing, labeling, and classification with Support Vector Machine (SVM) and Naïve Bayes. Data was collected from December 2023 to May 2024 with 1,790 reviews after going through preprocessing stages with algorithm Support Vector Machine and Naïve Bayes, with results accuracy respectively 82% and 81%. The results showed that Support Vector Machine's accuracy is taller than Naïve Bayes. Based on the negative sentiment, there are some problems with the quality of the service, namely service no can use, problems with feature applications, problems with system payments, problems with response to complaint users, and display issues with applications. Recommendations for improvements can be made, namely periodic evaluation and planning rapid recovery incidents, as well as optimization of application loading time. Monitor bait come back customer in a way routine and responsive with action right. Communicate with clear policy privacy and proactively provide service with a guide or automatic notification. This result gives an outlook important for Bank Mandiri in increasing the quality of service Livin ' by Mandiri.