Twitter Sentiment Analysis Using Recurrent Neural Network

Abstract

Being one of the largest social media platforms, Twitter has a diverse community collaborating on a multitude of ideas. With large amounts of data being generated and collected on Twitter every day, it is the perfect platform to implement machine learning algorithms to analyze information in different tweets. A recurrent Neural Network (RNN) is a specific algorithm that is used to solve problems involving sequential data such as texts or time series. In this paper, we use RNNs to classify a given tweet as either a positive or a negative sentiment. The RNN model we describe below was able to reach an accuracy of 76%.