PERBANDINGAN METODE SUPPORT VECTOR MACHINE DAN K-NEAREST NEIGHBOR DALAM ANALISIS SENTIMEN ULASAN APLIKASI MyXL

Abstract

Sentiment analysis is an essential tool for understanding user perceptions of mobile applications, especially when reviews are unstructured text. This study aims to analyze user reviews of the MyXL application using Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) and to compare their performance in classifying sentiments into positive, negative, and neutral categories. The dataset was obtained from Google Play Store via Kaggle and underwent text preprocessing, including case folding, removal of numbers and punctuation, tokenization, stopword removal, normalization, and stemming. Features were transformed into numerical representations using TF-IDF, and the data was split into training (70%) and testing (30%) sets. Evaluation using accuracy, precision, recall, and F1-score showed that SVM outperformed KNN with an accuracy of 0.743 versus 0.64, particularly in classifying neutral reviews. KNN exhibited higher misclassification in positive and negative classes, while SVM was more stable but tended to be biased toward the neutral class. These results provide insights for application developers to better understand user satisfaction and guide service improvement and feature development.