Sentiment Analysis of Bandung Tourist Destination Using Support Vector Machine and Naïve Bayes Algorithm

Bandung is the capital of West Java, also known as a famous tourist destination in Indonesia. Many places have been visited by many tourists. After their visit, they usually write their opinions about the place they visited on their favorite social network. People's opinions towards a particular object may contain any emotion, it can be positive or negative. Sentiment analysis is a process that analyzes people's opinions of the entity from the text pattern. Sentiment Analysis classifies an opinion into positive or negative sentiment using the Sentiment Analysis System. Analyzing people's sentiments about tourist destinations in English has been carried out by many researchers, but only a few in Indonesian. The aim of this study is to compare Support Vector Machine and Naïve Bayes performance using the synthetic Minority Oversampling Technique (SMOTE). SMOTE is a method used to handle data that is not balanced between classes. Support Vector Machine is an algorithm used to classify big data by maximizing margins using a hyperplane. And Naïve Bayes is a simple and efficient algorithm for handling classification data, Naïve Bayes has well performance for handling data with few features. The results show that the Support Vector Machine with SMOTE has a higher accuracy of 88.51% and 71.51% for Naive Bayes with SMOTE.