Comparison Of Sentiment Analysis Of Traveloka And Tiket.Com Applications On Twitter Using The Naive Bayes Method

Abstract

The country of Indonesia has a strategic geographical position and is also said to be a country that is very rich in natural resources and cultural diversity. One of the supporters of economic growth in Indonesia is tourism. To support the potential of the tourism sector in Indonesia, many online travel agent applications have started to appear. Of the many OTAs, the top two applications were selected, namely the Traveloka and Tiket.com applications. This sentiment analysis requires data from Twitter. This research compares sentiment analysis on the Traveloka and Tiket.com applications in terms of price and service. The method used is naïve Bayes. The goal is to get sentiment information contained in a text with a positive or negative view. With this research, it is hoped that we can see a comparison of sentiment analysis between the Traveloka and Tiket.com applications and be able to find out the level of accuracy of naïve bayes on the Traveloka and Tiket.com applications. The price dataset that gets more positive sentiment is the Traveloka price of 97.2%. In the service dataset that has positive sentiment, Tiket.com is 46.9%. Then, the greatest accuracy was obtained after oversampling the Tiket.com price dataset by 73%, Traveloka prices by 94%, Ticket services by 87% and Traveloka services by 86%.