Analisis Sentimen Komentar YouTube Terhadap OST Film Jumbo Menggunakan Metode Support Vector Machine (SVM)

Abstract

Sentiment analysis is a technique used to transform textual data into meaningful information by classifying opinions into positive, negative, and neutral categories. However, textual data often presents challenges such as unstructured characteristics and imbalanced class distribution, which can affect classification performance. This study aims to analyze the sentiment of YouTube comments on the Original Soundtrack (OST) of the Jumbo film entitled “Selalu Ada di Nadimu” using the SVM method. The research methods employed include collecting YouTube comment data, data preprocessing (cleaning, case folding, tokenization, normalization, stopword removal, and stemming), sentiment labeling using the InSet Lexicon, feature weighting using TF-IDF, and handling imbalanced data with the SMOTE method. Next, the data was divided into training and test sets, followed by classification using SVM with various kernels to obtain the best model. Model evaluation was conducted using a confusion matrix, including accuracy, precision, recall, and F1-score. The results show that the SVM method is capable of classifying sentiment into positive, negative, and neutral categories with good performance. The best model was achieved using the linear kernel with parameter C=100 and a data split of 90:10, resulting in an accuracy of 92,64%. In addition, the application of SMOTE improved model performance by balancing the class distribution, enabling the model to better recognize patterns in each sentiment category.