Komparasi Algoritma Support Vector Machine dengan Naive Bayes Untuk Analisis Sentimen Pada Aplikasi BRImo

Abstract

Banking is an industry that is currently developing in the use of information technology by increasing service quality standards in order to compete in the market in an increasingly tight digital era. At this time, BRI bank is attracting public attention to the quality of renewal by launching a mobile banking application, therefore an analysis of reviews of BRImo mobile banking users is carried out to serve as an object of research by comparing text mining classification methods. This study aims to determine the results of the comparison of the Support vector machine and Naive Bayes algorithms. Support vector machine and Naive Bayes algorithm are classification methods used to process data in the form of text with a good level of accuracy. This algorithm is usually used for text mining analysis with 4 stages, namely data scrapping, preprocessing, classification and evaluation. At the preprocessing stage, there are several processes including filtering, labeling, case folding, tokenization, stopword removal & stemming and normalization to get relevant words to be classified. The results of this study are the Support vector machine algorithm has a better performance in classifying the BRImo mobile banking application review data with an accuracy value of 97,69% compared to the Naive Baye algorithm with an accuracy value of 96,53%.