Analisis Sentimen Ulasan Netflix Menggunakan Model Deep Learning Berbasis Transformer (Arsitektur BERT)

Abstract

This study aims to evaluate the quality of Netflix's services through user reviews collected from the Google Playstore using text analytics techniques. The main approach used in this research is sentiment analysis based on the IndoBERT deep learning model. The research includes several stages, such as data collection, preprocessing, and classification into sentiment categories (positive, negative, and neutral). The result shows that sentiment analysis using pre-trained transformer-based models can provide valuable insights regarding user satisfaction and service quality perception. This study highlights how sentiment analysis applied to Playstore reviews can be a useful tool for companies to monitor user feedback and improve their service accordingly.