SENTIMENT ANALYSIS OF VIDEO EDITING APPLICATIONS USING SUPPORT VECTOR MACHINE ON GOOGLE COLAB

Abstract

Sentiment analysis is an important approach in understanding user opinions about an application. This study aims to analyze user reviews of the CapCut application using the Support Vector Machine (SVM) algorithm on the Google Colab platform. The preprocessing stages include data cleaning, word normalization using a dictionary from Kaggle, case folding, tokenization, stopword removal, and stemming. Furthermore, the data is converted into a numerical representation using the TF-IDF vectorization method. The labeling process is carried out using a sentiment lexicon obtained from GitHub. After performing data splitting, the SVM model is applied to classify sentiment into three categories: positive, negative, and neutral. The evaluation results show that the SVM model achieves the best accuracy of 90.12%. Based on the classification report, the model has high precision of 0.94 for positive and negative classes and 0.83 for the neutral class. Additionally, the confusion matrix indicates that the model can classify sentiment quite well, although there are still minor errors in predicting neutral sentiment. The findings of this study demonstrate that the SVM method can be effectively used to analyze user sentiment toward the CapCut application, providing valuable insights for improving user experience in the future.