Aspect Based Sentiment Analysis pada Ulasan Pengguna Aplikasi Pedulilindungi Menggunakan Support Vector Machine

Abstract

Various efforts were made by the government of the Republic of Indonesia to deal with the spread of the COVID-19 virus in Indonesia, one of the breakthroughs from the government through the Ministry of Communication and Information (KOMINFO), the Ministry of Health, the Ministry of BUMN and the National Disaster Management Agency was to create the PeduliLindungi application. PeduliLindungi is an application developed to stop the transmission of Coronavirus Disease (COVID-19). As an application that is needed according to IT experts, the "PeduliLindungi" application still has many problems that cause user dissatisfaction and if it is not fixed immediately user interest will decrease, so the development process is very necessary. Development requires opinions from users and the data can be obtained using google play store reviews about the PeduliLindungi application. However, in the process of identifying these reviews, it is still necessary to carry out sentiment analysis based on the aspects contained in the review sentence. The purpose of this study was to analyze sentiment based on aspects of user reviews of the PeduliLindungi application using the SVM (Support Vector Machine) method. This study uses 1022 data, with the application of feature extraction in the form of TF-IDF and will be identified using the Support Vector Machine method. By applying the confusion matrix evaluation method, this study obtained a total average accuracy score based on four aspects of 91%.