<![CDATA[The popularity of diet coffee, particularly the Americano variant, has risen alongside the growing trend of healthy lifestyles in society. This phenomenon has led to various public opinions circulating on social media, which need to be analyzed to better understand consumer perceptions. This study compares two commonly used text feature representation methods, Term Frequency-Inverse Document Frequency (TF-IDF) and Bag of Words (BoW), in sentiment analysis using the Naïve Bayes algorithm. Using relevant keywords, data were collected from Twitter and underwent preprocessing stages including case folding, cleansing, tokenizing, stopword removal, and stemming. Sentiment labeling was conducted manually based on keyword indicators, and the data were classified into positive, negative, and neutral categories. The evaluation results show that the TF-IDF model achieved an accuracy of 85%, outperforming BoW which obtained 64%. This performance gap indicates that the choice of feature representation method plays a crucial role in the success of sentiment classification. This research is expected to serve as a reference for optimizing text representation techniques to analyze public opinion on social media, particularly concerning diet products and low-calorie beverages.]]>
Copyrights © 2025
