Analisis Sentimen Terhadap Makanan Manis di Platform X Menggunakan TF-IDF dan Naive Bayes

Abstract

The rise of social media has made it easier for people to share opinions on various topics, including sweet foods. Sentiment analysis of these opinions provides valuable insights for the food industry to understand consumer perceptions, identify preferences, and design marketing strategies. This study aims to analyze sentiment in reviews of sweet foods using the Naive Bayes algorithm. The data used was collected from platform X, consisting of reviews in Indonesian. The analysis process includes data cleaning, tokenization, stopword removal, stemming, and feature extraction using TF-IDF. Sentiments are classified into three categories: positive, negative, and neutral. The results show that the Naive Bayes algorithm achieved an accuracy of 67.1%, with significant contributions in classifying positive sentiments as the dominant category. These findings provide essential insights for food producers to understand consumer perceptions of sweet foods and serve as a reference for research in sentiment analysis and digital marketing.