<![CDATA[The large number of consumer reviews of skincare products such as Skintific 5X Ceramide Barrier Repair Moisture Gel on e-commerce platforms raises the need for automated sentiment analysis. This study classifies 2,000 user comments from the Sociolla app using the Support Vector Machine (SVM) algorithm. Data were obtained through web scraping and processed through preprocessing, lexicon-based labeling, and word weighting using TF-IDF. SVM with a linear kernel was used to distinguish positive and negative comments. Performance evaluation using a confusion matrix resulted in an accuracy of 89.73%, a precision of 0.94, a recall of 0.75, and an F1-score of 0.83 for the positive class, and a precision of 0.88, a recall of 0.98, and an F1-score of 0.93 for the negative class. These results indicate that SVM is effective for sentiment classification in online beauty product reviews.Keywords: E-commerce; Support Vector Machine; Sentiment Analysis; TF-IDF; Sociolla; Skintific; Lexicon-based AbstrakBanyaknya ulasan konsumen terhadap produk perawatan kulit seperti Skintific 5X Ceramide Barrier Repair Moisture Gel di platform e-commerce menimbulkan kebutuhan akan analisis sentimen otomatis. Penelitian ini mengklasifikasikan 2000 komentar pengguna dari aplikasi Sociolla menggunakan algoritma Support Vector Machine (SVM). Data diperoleh melalui web scraping dan diproses dengan tahapan preprocessing, pelabelan berbasis lexicon, serta pembobotan kata menggunakan TF-IDF. SVM dengan linear kernel digunakan untuk membedakan komentar positif dan negatif. Evaluasi performa menggunakan confusion matrix menghasilkan akurasi sebesar 89,73%, precision 0,94, recall 0,75, dan F1-score 0,83 untuk kelas positif, serta precision 0,88, recall 0,98, dan F1-score 0,93 untuk kelas negatif. Hasil ini menunjukkan bahwa SVM efektif untuk klasifikasi sentimen pada ulasan produk kecantikan secara daring. ]]>
