ANALISIS SENTIMEN KEPUASAN PELANGGAN TRANSPORTASI ONLINE PADA GRAB MENGGUNAKAN SUPPORT VERCTOR MACHINE

Abstrak

Online transportation services such as Grab have become an essential part of urban mobility in Indonesia, generating a wide range of user reviews that reflect levels of satisfaction. This study aims to analyze the sentiment of these reviews using the Support Vector Machine (SVM) algorithm. Data were collected from the Google Play Store and processed through several stages, including text preprocessing, automatic labeling based on rating scores (≤3 as negative, ≥4 as positive), and feature representation using the Term Frequency–Inverse Document Frequency (TF-IDF) method. The dataset was split into training data (80%) and testing data (20%), and the SVM model was trained using a linear kernel. Evaluation results showed an accuracy of 82%, precision of 84%, recall of 78%, F1-score of 79%, and an AUC of 0.9015. Further analysis of negative reviews revealed that the aspects of “drivers,” “application,” and “payment” were the main sources of complaints. These findings demonstrate the effectiveness of SVM in sentiment classification and its potential as a data-driven service evaluation tool. The study also recommends manual labeling or semantic-based approaches to address inconsistencies between review scores and content.