Stock Market Prediction Using Machine Learning

Abstract

Predicting stock prices accurately is very important in ever-changing political and economic situations in order to minimize risks and maximize returns.Due to the high rate of returns, liquidity, and dividends investors prefer to invest in stocks. Doing research on stocks is a never-ending phenomenon. It is an ever-green field. Due to inefficiency in the market, people have been looking for ways to earn more profit by using computational methods, machine learning, mathematical models, and algorithms. Researchers found that time series method is the best method for predicting stock prices. Due to non linearity of data deep learning methods such as LSTM, ANN, RNN and ARMA and ARIMA were used because they cannot only are able to process nonlinear data but are also able to retain specific sequences in their memory. In this article, we will discuss time series method and different types of neural networks to predict stock prices. Accuracy will be measured by mean square error of different models.Since LSTM can handle nonlinear data better than time series method so we will use LSTM in our project.