Analisis Sentimen Ulasan Aplikasi Citilink Menggunakan Metode Support Vector Machine dengan TF-IDF

Abstract

In line with the advancement of the Industry 4.0 era, Indonesian society has been living side by side and is inseparable from the existing technological advancements. One of the conveniences experienced by today's society is that transactions no longer need to be conducted face-to-face in a particular place but can now be done online. In the context of air transportation, technological advancements have been very helpful to the public. Airline applications are one of the most widely used by passengers. In this study, the researchers focused on analyzing public sentiment towards the Citilink application, one of Indonesia's leading airlines. The researchers used the Support Vector Machine (SVM) method enhanced with TF-IDF (Term Frequency-Inverse Document Frequency) text representation to analyze sentiment from user reviews. The stages of this research began with data collection containing reviews from the Citilink application to analyze its sentiment. Then, it proceeded to the data preprocessing stage, where the collected data was cleaned until it became tokens ready for testing. After that, it moved to the weighting stage using Term Frequency-Inverse Document Frequency (TF-IDF). Then it continued to the stage of applying the Support Vector Machine (SVM) model. The last one is the evaluation to measure the accuracy level of the model used. Based on the results of this study, it can be concluded that the Support Vector Machine model that has been adapted to the dataset of Citilink application reviews from Google Playstore and supported by TF-IDF feature extraction successfully classified the sentiment of reviews with high accuracy, reaching 88%. Further evaluation also showed satisfactory values of precision, recall, and F1-Score, namely 90%, 83%, and 85%, respectively. This study shows that the Support Vector Machine model can be an effective instrument in understanding user responses to the performance of the Citilink application.