Instagram Sentiment Analysis with Naive Bayes and KNN: Exploring Customer Satisfaction of Digital Payment Services in Indonesia

Abstract

Internet penetration and trends of social media have generated a load of data and many researchers have done sentiment analysis to know user's customer satisfaction of a particular brand through textual data form. This research was conducted to determine the customer satisfaction of digital payment services (Go-Pay, Ovo, and LinkAja) in Indonesia, analyzed from the result of Instagram sentiment analysis, in which the textual data form used is the Instagram comment. The research model was built with sentiment analysis classification algorithm: Naïve Bayes and K-Nearest Neighbors (KNN), and customer satisfaction theory. The result of this research with 3800 training data and 200 test data by using 20 k-fold cross validation shows that GO-PAY has all kinds of customers who satisfied, not satisfied and both. As for OVO and LinkAja have a more neutral sentiment. The most accurate classification algorithm implemented in this research is the KNN classifier algorithm, up to 90%. The final sentiment analysis result of this research contributed to the current state of customer satisfaction towards Go-Pay, Ovo, and LinkAja as a digital payment service provider in Indonesia and maybe an input to the relevant stakeholder.