Stock Price Prediction using LSTM-ARIMA Hybrid Neural Network Model with Sentiment Analysis of News Headlines

Financial markets are extremely volatile, which ends in people losing their money within the exchange. Our project- Stock Price Prediction using LSTM-ARIMA Hybrid Neural Network Model with Sentiment Analysis of News Headlines, is one amongst the many approaches to unravel the matter and predict accurate stock prices. Due to the noise and volatility of the stock market, timely market prediction is typically regarded as one of the most difficult challenges. We suggest a deep learning-based stock market prediction model that takes investors’ emotional tendencies into account to overcome these issues. This paper uses a unique method to predict next day’s final stock prices using a combination of LSTM, ARIMA statistical model and Sentiment analysis. This project will mainly provide an insight to traders and investors about future stock prices, thus helping them make the right decisions. They are going to thus be able to minimize the loss of their money and resources.

I. Introduction

Stock prediction is of extreme importance to banking, investment and stock exchange firms. A huge amount of stock price data is thus used for the prediction of stock prices. Stock values, are affected by a lot of factors like news, opening price, closing price, etc.