Opinion Mining and Sentiment Analysis on Twitter Data

Sentiment analysis on Twitter data provides organizations with a quick and effective method to monitor the general public's sentiments about a brand, business, leaders, and so on. In recent years, a wide range of parameters and techniques for training sentiment analyzers for Twitter datasets has been researched, with various outcomes. To perform sentiment analysis on tweets, our proposed system extracts the posts from Twitter and analyzes tweet's sentiments towards products and services. In this paper, we used Long-Short Term Memory (a type of recurrent neural network) for Text Classification as it effectively enhances performance by memorizing vital information and detecting patterns. Using LSTM, we implemented an algorithm for automatic classification of text/tweets into six different sentiment classes, which can be extended further depending on the requirements. Our results showed a precision of 92.17% on testing the model based on test sequences and test labels and we visually expressed the sentiments in the form of a Pie-Chart.