SENTIMENT ANALYSIS ON ONLINE REVIEWS USING NAÃVE BAYES CLASSIFIER
ABSTRACT
The Increasing number of population, the growing of information technology, and communication trigger the
increasing of economic activity of society. Increased economic community cannot be separated from the process of ongoing
transportation. Transportation is a service that is needed every day for the community. The increasing number of means of
transportation has an impact on the intense competition between these types of businesses. The level of customer satisfaction is a
vital aspect to survive and win business competition, one way of measuring customer satisfaction is by providing easy and
convenient access for customers to submit suggestions, criticisms, opinions, and complaints. Along with the development of online
media, customer reviews can be created and viewed by many people through social media. This study uses web scraping
techniques to obtain Garuda Indonesia airline reviews data from the TripAdvisor site. Data obtained from the TripAdvisor site is
further labeled and analyzed using the Naïve Bayes Classifier (NBC) method to classify reviews based on positive and negative
sentiment categories. Furthermore, the results of sentiment classification will be analyzed by Text Mining method, the main
concept is to do the widest exploration in the data of the reviews that have been obtained so that it is found an information that is
considered important and can be useful for various areas of need. Sentiment classification results show more than 80% of the
reviews are positive reviews with an accuracy of 82.02%. From the text association results obtained information that passengers
of Garuda Indonesia majority share talk about the service, staff, seat, and food because it always appears in both positive and
negative sentiment class.
