Stock Prediction Using Twitter Sentiment Analysis

Prediction and analysis of stock market data are very important in today’s day and age. Since the economic interactions are too complex for shallow neural networks this paper implements Long Short Term Memory (LSTM) neural networks. LSTM is chosen as it helps to vectorize the data and thus give better predictions. This paper agrees that longer horizon predictions e.g. a month are more useful than shorter horizon e.g. a day. A very important factor is the mood of the people. A person’s emotions have the power to influence the stock market. Sentiment analysis on twitter is used to find a correlation between the future of the stock and the general public’s mood. Our paper works on comparing the sentiment analysis and the predicted stock value and showing that the two are rather similar and that people’s emotions affects the future of the stock prices and to do a comp