Analisis Sentimen Berbasis Aspek Pada Aplikasi Polri Super App Dengan Support Vector Machine Dan Word2vec

Abstract

The police super app is a form of police transformation in providing one-stop public services to the community. Through this application, the public can access various police services. Although this application allows the public to access various police services, there are still some weaknesses. The purpose of this research is to evaluate the Polri Super App application through reviews in the play store using the Support Vector Machine and word2vec methods in sentiment analysis that focus on certain aspects and their performance. The results of research with various tests show that the best model is obtained in SVM Sentiment and Aspects without SMOTE with Sentiment accuracy at 90% and aspects at 75 to 80%.