Sentimental Analysis Using RNN, CNN AND LSTM: A Comparative Study Of Accuracy And Computational Efficiency

Abstract

With the growing trend of the social media, people express their emotions on the different platforms in the form of comments or reviews whether it is positive, negative or neutral. Social networking sites like Twitter, instagram, youtube etc. are gaining popularity very quickly as they allow people to express and share their thoughts on topics, communicate with different communities or spread it to the world. Many studies have been done in the field of data on social media. In political elections, it is used to track political views and detect inconsistencies and discrepancies between talk and action at the government level. It can also be used to predict election results. Now a days election polling is done to predict the winning party and which party’s manifesto is grabbing the public interest.

In this paper, different models such as CNN, RNN and LSTM are used in order to compute accuracy and computation time. The work is done on the real-time data.