Analisis Sentimen Ulasan Aplikasi Transportasi Online Menggunakan Multinomial Naïve Bayes dan Query Expansion Ranking

Abstract

The rapid development of the transportation industry in recent years has led to a new innovation in the field of transportation, namely the application of online transportation services. To facilitate the translation of user satisfaction, in addition to users being able to provide reviews, the Google Play Store uses a rating system consisting of a rating of 1 to 5. However, users often do not provide a rating that is in accordance with the review so that this is not enough to determine the sentiment of the review. This research is focused on evaluating the performance of the features selection using Query Expanison Ranking on the Multinomial Naïve Bayes method in the problem of sentiment analysis on the two of most popular online transportation service applications in Indonesia, namely Gojek and Grab. From the results of the performance evaluation using k-fold cross validation, it was found that the best feature selection ratio was 20% with the best performance in terms of precision.