Sentiment Analysis on IMDB Movie Reviews using BERT

Before technology existed, opinions could only be obtained from acquaintances, friends, or experts who were experts in certain fields. However, as technology develops, it turns out that opinions can be expressed through social media so that they can influence everyone who sees them. One of them is movie reviews. Human opinion about something is often not valid. So, this study aims to investigate the sentiment analysis related to IMDB Movie Reviews. The approach used is BERT. BERT is a deep learning approach. The data used in this study is the IMDB Movie Review of 50,000 data. The existing data is divided into three parts, namely training data, validation data, and testing data. The results obtained from the BERT model are 91.69% for training accuracy 0.187 for training loss, 91.85% for validation accuracy, 0.212 for validation loss, 91.78% for testing accuracy, and 0.207 for testing loss. It can be seen, that BERT is a very effective approach for sentiment analysis of IMDB Movie Review so that the research problem regarding the invalidity of one's opinion can be handled properly.