Analisis Sentimen Aplikasi MyXL menggunakan Metode Support Vector Machine berdasarkan Ulasan Pengguna di Google Play Store
Kata Kunci:
Sentimen Analisis, Scraping, Text Preprocessing, TF-IDF, Support Vector Machine, Google Play StoreAbstrak
Review on Google Play Store is one of the features used to provide a rating of an application. MyXL is a self-service application provided by PT XL Axiata Tbk on the Google Play Store that is useful for users to perform XL services easier such as activating internet packages, checking credit, checking remaining quota, etc. However, the review on the application is only in the form of text with no specific meaning and there are some high ratings but the reviews given are negative reviews, for that a sentiment analysis is needed that can classify reviews as user sentiment. In this research, the scraping stage was carried out for collecting application user review data, followed by the text preprocessing stage to process data by selecting data and turning it into more structured data. The data from the text preprocessing were word weighted using the Term Frequency - Inverse Document Frequency (TF-IDF) method. Then sentiment classification is carried out using the Support Vector Machine (SVM) algorithm. The best results were obtained with the SVM algorithm for sentiment testing for 2 classes using the value of training data and test data of 80%:20%, the total data is balanced with 160 positive and 160 negative data, experiments with cross validation K = 5 and the use of a linear kernel. The results obtained for the average value of 88% accuracy, 88% precision, 88% recall and 88% f-measure.