<![CDATA[The growth of the tourism sector in Palembang City has encouraged an increase in the need for quality hospitality services. In the digital age, user reviews on the Google Maps platform are an important source of data to assess customer satisfaction. This study aims to analyze and compare the effectiveness of two sentiment classification algorithms, namely Support Vector Machine (SVM) and Random Forest, in processing hotel reviews in Indonesian. A total of 1000 review data was used and processed through the stages of text cleanup, letter normalization, tokenization, stopword removal, and stemming. The evaluation was carried out with two approaches: 80:20 data sharing and cross-validation using the K-Fold technique. On data sharing, Random Forest showed 88% accuracy and 100% recall, while SVM recorded 87% accuracy and 99% recall, with equivalent precision and F1-score. However, cross-validation showed that the SVM was more stable and consistent, with 92% accuracy, 94% accuracy, 98% recall, and 96% F1-score, outpacing Random Forest's 91% accuracy and 95% F1-score. These results show that the SVM algorithm is superior in analyzing hotel review sentiment on Google Maps. These findings provide recommendations for tourism information system developers to adopt an SVM-based approach to review data processing to support more accurate and responsive decision-making.]]>
Copyrights © 2025
