SENTIMENT ANALYSIS OF BAG BRANDS ON PLATFORM X (TWITTER) USING NAIVE BAYES CLASSIFICATION METHOD AND SUPPORT VECTOR MACHINE (SVM)

Abstract

The development of social media has changed the way consumers express their opinions and experiences regarding a product. Platform X (Twitter) has become one of the most actively used channels for expressing opinions, both positive and negative, about a brand. As a fashion product, handbags are closely linked to consumers' perceptions of design, quality, price, and lifestyle, making it crucial for companies to understand public sentiment in order to develop effective marketing strategies.

This study aims to investigate public perceptions and sentiments toward handbag brands through an analysis of user posts on Platform X (Twitter). The analysis was conducted by comparing the performance of two text classification methods, namely Naive Bayes and Support Vector Machine (SVM), to determine positive, negative, and neutral sentiments. By understanding sentiment patterns, it is hoped that useful insights can be obtained in designing more targeted promotional and product development strategies.

The research data was obtained through a process of web scraping tweets containing keywords relevant to the bag brand. The collected data was then processed through preprocessing stages, including case folding, tokenization, stopword removal, and stemming. The vectorization process was performed using the Term Frequency–Inverse Document Frequency (TF-IDF) method before being applied to the Naive Bayes and SVM models for classification. Model performance evaluation was conducted using the Accuracy, Precision, Recall, and F1-Score metrics.

The results of the study show that the SVM method produces better performance than Naive Bayes, especially in terms of accuracy and precision in identifying positive and neutral sentiments. The accuracy percentage obtained by SVM is at a higher level, indicating its ability to separate text data that has similar characteristics but different sentiments. This shows that SVM is more effective when used on complex text data with high language variation, such as that found on the X (Twitter) platform.

Based on these findings, it can be concluded that sentiment analysis using the SVM method can be a more accurate alternative for mapping public opinion on bag brands on social media. The results of this study can be used by companies as a basis for strategic decision-making, whether in marketing campaigns, product innovation, or service quality improvement. Additionally, this research opens opportunities for the development of automated systems to monitor consumer sentiment in real-time, enabling companies to be more responsive to market dynamics.