Predictive Analysis on Stock Market Data using Sentiment Analysis and Long Short-Term Memory (LSTM) Neural Network

Author: Siddharth Chauhan
Chair: Erik Golen - efgics
Committee: Dave Patric - dkpvcs
Submitted On: Jan 1, 2021
Machine Learning

Stock market prediction has been one of the most interesting use cases of machine learning for a long time. Stock market prediction involves both fundamental and technical analysis of a particular stock. Fundamental analysis of stocks itself depends on multiple factors – physical factors, behavioral economics, news about a particular stock, demonetization, monetary policy, natural disaster, etc. Existing studies have shown that there has been a strong correlation between news article, blogs, and stock price momentum. The project aims to use text mining (Natural Language techniques) to predict stock market momentum and time-series data to further improve the accuracy of prediction. The time series model will use Long Short-Term Memory (LSTM) neural network to predict future 30 days stock close price which will cover the technical analysis of stock momentum. The hybrid approach of using both sentiment analysis and the LSTM model will result in high predictability of the stock market.