Stock Price Prediction Using ARIMA, LR and LSTM

Stock price prediction is a necessary topic in economics finance which has ignited the interest of various researchers over the decade to research for better predictive models. The model ARIMA i.e., Auto Regressive Integrated Moving Average, model LSTM i.e., Long Short Term Memory and the LR, i.e. Linear Regression model for the prediction of Stock prices for the current time period is on the trend for use and one of the best models of course. The model uses synthetic dataset that is a real time dataset from yahoo finance website. The models will work on the stock prices of Bank of America. This study evaluates and compares both the models using the parameters like Root Mean Square Error (RMSE) and Mean Absolute Percentage error (MAPE). The study further analyzes models for accuracy obtained for prediction of stock price. After comparison, this study suggests that even though all three models can predict stock prices to a good amount of accuracy, the proposed model using LSTM out performs and provides better accuracy with lower error rate.