<![CDATA[The advancement of information and communication technology has transformed consumer interactions with products and brands, especially in the beauty industry. This study focuses on sentiment analysis of sunscreen product reviews using the Naive Bayes Classifier method. Review data for the Wardah UV Shield Essential Sunscreen Gel SPF 35 PA+++ were collected through web scraping from the Femaledaily website, resulting in 1,451 data entries. The data were labeled as positive or negative based on ratings and then processed through data cleaning, case folding, stopword removal, and tokenization. The processed data were converted into numerical representations using TF-IDF. The Naive Bayes Classifier model built for this study achieved an accuracy of 79%, precision of 67%, recall of 64%, and an F1-score of 65%. A word cloud visualization highlighted frequently occurring words in both positive and negative reviews. This study demonstrates that the Naive Bayes Classifier method is effective for classifying sentiments in sunscreen product reviews. Although this method is easy to implement and understand, it has limitations due to the assumption of word independence and the imbalance between positive and negative reviews. Future research is expected to expand the dataset and explore other sentiment analysis methods to improve accuracy.]]>
Copyrights © 2025
