<![CDATA[The development of e-commerce in Indonesia, particularly Tokopedia, has led to an increase in the number of reviews given by consumers after making a purchase. These reviews consist of positive and negative opinions that can be used to assess product quality. However, the high volume of reviews makes manual analysis ineffective. This study aims to classify product review sentiment on Tokopedia using the Naïve Bayes algorithm to determine the best product quality from several stores selling similar items. The Naïve Bayes algorithm was chosen because of its good classification capabilities and efficiency in text data processing. The results show that Naïve Bayes is capable of classifying reviews as positive or negative with an average accuracy of 95% on the test data. Testing per store also yielded consistent results with an accuracy of around 90%. Furthermore, the sentiment results were converted into quality metrics with the categories Good, Fair, and Poor. Products at Store A scored 85% and Store B scored 80%, so both were categorized as Good, with Store A being more recommended. This study shows that Naïve Bayes-based sentiment analysis is effective for assessing product quality and provides insights for sellers to improve their services.]]>
Copyrights © 2025
