ANALISIS SENTIMEN PENGGUNA TWITTER TERHADAP SKINCARE DENGAN METODE SUPPORT VECTOR MACHINE (SVM)

Abstract

The Originote Hyalucera Moisturizer skincare product has attracted public attention because it offers superior quality at an affordable price. Social media, especially Twitter, is used by consumers to express opinions regarding this product, whether positive, negative, or neutral. However, the large number of reviews with various sentiments can confuse potential consumers in assessing product quality. Therefore, this study aims to understand user perception through sentiment analysis and evaluate the effectiveness of the Support Vector Machine (SVM) algorithm in sentiment classification. A total of 1,820 tweets were collected using the crawling technique with Python. The data undergoes preprocessing, including text cleaning, tokenization, stopword removal, and stemming, reducing it to 902 tweets. Key text features are extracted using Term Frequency-Inverse Document Frequency (TF-IDF). For sentiment classification, this study used the SVM algorithm, which is known as an effective method in text processing. Model evaluation showed good results with an accuracy of 87%, precision of 89%, and recall of 87%. This study provides insight into public perception of The Originote Hyalucera Moisturizer and measures the effectiveness of SVM in social media-based sentiment analysis. The results of the study can be utilized by manufacturers for more targeted marketing strategies, product quality improvement, and more effective communication in responding to opinions on social media. In addition, this study contributes to the development of machine learning-based sentiment analysis methods in the context of skincare products.