Predicting Customer Sentiment in E-Commerce: Leveraging Naive Bayes and Support Vector Machine Models Using the Twitter API

Sentiment analysis also known as opinion mining, is a technique to identify, extract, and evaluate the subjective data in texts using natural language processing. In the digital era, the widespread use of the internet to share opinions through online reviews offers a wealth of sentiment data that is important for public opinion analysis, understanding consumer expectations and preferences. This research focuses exclusively on Twitter comments as the primary data source. Popular sentiment analysis techniques include Support Vector Machine (SVM), and Naïve Bayes are used to categorize reviews and comments into positive, negative, or neutral sentiment. From the experiments that have been carried out, the SVM model on the Tik Tok Shop dataset has an accuracy of 80%, on the Zalora dataset has an accuracy of 75%, and the Shopee dataset has an accuracy of 90%, then for the NB model on the Tiktok Shop dataset has an accuracy of 70%, on the Zalora dataset has an accuracy of 83%, and the Shopee dataset it has an accuracy of 90%. Experimental results show that the SVM model can effectively classify comments with higher accuracy than Naive Bayes.