Application of ARIMA and LSTM Model to the Forecast of CSI 300 Close Price

Abstract

This article uses two time-series analysis model, ARIMA model and Long Short-Term Memory neuron network model to predict the close price of CSI 300 Index. The ARIMA model only uses the past close price values to predict future close price while the LSTM model combine multiple variables into the prediction of future close price. The methods of fitting data with these two models are introduced. The predictive performance of the two fitted models is evaluated by calculating the Root Mean Squared Error, which is an indicator of the difference between predicted values and real values. This article then analyzes the prediction results and proposes the respective advantages and limitations of the two models. It is found that the ARIMA model performs better when predicting close price in a shorter period but may not perform well when doing long-term prediction. The LSTM model can predict a similar series trend with multiple variables and its parameters can be adjusted to get smaller RMSE and improve prediction performance.