Sentiment analysis on customer satisfaction of digital payment in Indonesia: A comparative study using KNN and Naïve Bayes

Abstract

Abstract Indonesia payment behaviour has turned from traditional to digital as the impact of the technology growth. Digital payment usage in Indonesia have increased rapidly in recent years. Many companies offer this service with different terms and tariff options. Social media is one of the places where people can express their feeling and opinion, including Twitter. In this research, sentiment analysis and opinion mining is conducted to see public satisfaction towards the digital payment service in Indonesia (OVO, GO-PAY and LinkAja). This research is using Twitter data and has several stages, which are data crawling from Twitter, data cleaning, feature selection and classification using two machine learning approach (Naïve bayes classifier and K-Nearest Neighbour or KNN). The raw data is processed to get the clean data, and to get the appropriate feature for classification algorithm and then perform classification and validation to the model. As for the classification algorithm, this research finds out that KNN has better accuracy than Naive Bayes. The result of this research also shows that LinkAja and G0-PAY has more neutral sentiment or customers nearly satisfied of the services provided, and OVO has more negative sentiment than neutral sentiment