Analisis Sentimen Google Maps Review Wulan Rent Car Menggunakan Support Vector Machine (SVM)

Abstract

Google Maps reviews serve as a vital data source for companies including car rental businesses to understand consumer perceptions. Manually evaluating such vast amounts of data is inefficient; therefore, technology-based approaches are required to analyze the emotions or opinions embedded in the text (sentiment analysis). This study aims to classify the sentiment of Google Maps reviews for Wulan Rent Car using the Support Vector Machine (SVM) method. Out of 443 total reviews, 301 contained text and underwent further processing, including labeling via the Bing Liu Opinion Lexicon, data pre-processing, and TF-IDF-based scoring. The labeling results identified 269 positive, 18 neutral, and 14 negative reviews. Model evaluation yielded an accuracy of 84.75%. Based on precision, recall, and F1-score metrics, the model demonstrated strong performance for the positive sentiment class but struggled to optimally classify negative sentiment due to imbalanced data distribution. The findings indicate that while the SVM method applies to sentiment analysis of Wulan Rent Car’s Google Maps reviews, model performance remains influenced by data distribution characteristics, particularly regarding minority classes. Consequently, future research should incorporate data balancing techniques—such as the Synthetic Minority Over-Sampling Technique (SMOTE) or similar methods—to enhance the model's ability to classify all sentiment categories effectively.