Analisis Sentimen Pengguna Terhadap Game Minecraft pada Ulasan Google Play Menggunakan Metode Naïve Bayes dan Support Vector Machine (SVM)

Abstract

User reviews on digital platforms are an important source of information for understanding perceptions of game products, but the large amount of data makes manual analysis less effective. This study aims to analyze user sentiment towards the Minecraft game based on Google Play Store reviews and compare the performance of the Naïve Bayes and Support Vector Machine (SVM) methods. The study used a quantitative approach with 1,600 scraped reviews, which after preprocessing became 1,477 data. The research stages included text cleaning, sentiment labeling, TF-IDF feature extraction, data division using a stratified split, classification model development, and evaluation using accuracy, precision, recall, F1-score, and the McNemar test. The results showed that 70.3% of reviews had positive sentiment, 23.3% were negative, and 6.4% were neutral. The SVM model produced better performance with an accuracy of 82.09% and an F1-score of 0.79, compared to Naïve Bayes with an accuracy of 75.62% and an F1-score of 0.6788. However, the McNemar test results showed no significant difference in performance between the two models (p-value = 0.3074). This finding suggests that the SVM is superior in predictive performance, but both methods have relatively equivalent classification capabilities statistically in analyzing Minecraft user review sentiment.