Sentiment Analysis of IMDB Movies Dataset Using Deep Learning Based GRU Model

Sentiment Analysis is the need of the hour for every big or small companies, be it manufacturing or entertainment industries, educational institutions, media houses, political parties or even individual. All organizations want to identify churn and to get their customers retained with them by mean of sentiment analysis on social media. Increasing number of Internet and social media users contribute in petabytes of unstructured data generated every day. The continuous advancement in technology sector, in terms of high storage capability, computational processing power of the computers, cloud technology etc., has enabled processing of these high-volume unstructured data with ease, in a very less time, which otherwise would have taken even days to process. Many sentiment analysis systems are modelled by using deep learning, by using GRU (gate recurrent unit) architecture, has showed significant improvements with high tendency to reveal the underlying semantic meaning in the input text. GRU model seems to perform even better with lowest loss 0.0783, variable loss 0.4728, highest accuracy 97.20% and variable accuracy 88.19%.