Performance Analysis of Different Neural Networks for Sentiment Analysis on IMDb Movie Reviews

Abstract

With the huge expansion of text data sentiment analysis is playing a crucial role in analyzing the user’s perspective about a particular product, company or any other physical or virtual entity. Sentiment analysis helps us to analyze user review about an entity and then drawing out a conclusion based on the sentiments it extracted from the reviews. Convolution Neural Network (CNN) and Long-Short-Term Memory Network (LSTM) are two well-known deep neural networks used for sentiment analysis. In this paper, we have compared between CNN, LSTM and LSTM-CNN architectures for sentiment classification on the IMDb movie reviews in order to find the best-suited architecture for the dataset. Experimental results have shown that CNN has achieved an F-Score of 91% which has outperformed LSTM, LSTM-CNN and other state-of-the-art approaches for sentiment classification on IMDb movie reviews.