Analisis Sentimen Terhadap Produk Skincare Menggunakan Metode CNN, RNN dan SVM

Abstract

Sentiment analysis is an essential process for understanding user perceptions of a product, especially in digital platforms such as social media. This study aims to classify sentiment in user comments about skincare products—particularly Somethinc—into three categories: positive, negative, and neutral. The data used consists of user-generated comments that have undergone text preprocessing steps, including case folding, tokenization, removal of special characters, normalization, stopword removal, and stemming. Three classification methods were applied in this research: Support Vector Machine (SVM), Convolutional Neural Network (CNN), and Recurrent Neural Network (RNN). Based on evaluation results, the SVM method achieved the highest accuracy at 94%, followed by RNN at 86% and CNN at 85%. These results indicate that machine learning and deep learning approaches are effective in classifying public opinions about skincare products and can serve as a reference for producers to automatically understand consumer responses.