Analisis Sentimen Ulasan Pengguna Aplikasi E-Commerce Toco Menggunakan Algoritma Naive Bayes

This study aims to analyze the sentiment of user reviews of the Toco e-commerce application using the Naïve Bayes Multinomial algorithm. The dataset consists of 1425 reviews with an unbalanced distribution between positive and negative classes. Review data was collected from the Google Play Store platform, then processed automatically through the stages of case folding, normalization, stopword removal, and stemming. Modeling was carried out by dividing the data into training and test data, and classifying sentiment using the Naïve Bayes approach. From the evaluation results, the model's accuracy in sentiment classification reached 88%, with higher performance achieved in the majority class (positive) compared to the minority class (negative), as reflected in the low precision and recall values. This study emphasizes the need to handle unbalanced data so that the analysis results reflect the diverse perceptions of users. This research provides a baseline for sentiment analysis for local e-commerce applications and contributes to the development of automated analytics systems to support decision-making in the Indonesian e-commerce industry.