Twitter Sentiment Analysis using Natural Language Processing

Social media is the richest source of text generated by the user. So there is a necessity to automate the system to help organizing and classifying the opinions posted on social media sites. Proposed methodology framework using Artificial Recurrent Neural Network (ARNN) with bi-directional long short term memory (LSTM) has been used for the classification of sentiments. Structure for RNN with bidirectional LSTM is depicted. US airline Twitter sentiment dataset has been analysed using bidirectional LSTM model. Text with varying length is taken for the experiment. Graphical representation of the analysis has been depicted in this paper. Confusion matrix shows the result. At the end it is concluded that the sentiments are analysed and classified as positive, negative or neutral.