The Prediction Of CSI 300 Stock Index

The stock market,which is closely related to the real economy,can reflect the operation of the national economy.The stock index can reflect the overall trend of the stock market,affect the investor's nerve.The big boom and slump in the stock market will have a major impact on the real economy,and bring huge losses to the investors.Therefore,the prediction of the stock index has profound theoretical and practical significance.This paper find the shortcomings of single models and believe the combined model is the main direction of future research.Firstly,This paper forecasts the daily closing price of CSI 300 by using two single forecasting models-ARIMA model and grey artificial neural network model.The ARIMA model is modeled and predicted using stock index closing price time series.The GRANN model takes into account the influence of technical indicators on the closing price of the stock index: we choose 15 technical indexes related to the closing price of the stock index,then the grey correlation analysis is used to select the 8 variables with the highest correlation of closing price of the next day,which is used as input variable of artificial neural network,the final prediction results are obtained by setting the parameter training model.The empirical results show that the grey artificial neural network outperforms the ARIMA model.This paper integrates the two single models of ARIMA and GRANN,and proposes to use the ARIMA-ANN combination model and the GRANN-ARIMA combination model to predict the closing price of CSI 300.The hybrid sequence in the conventional hybrid is normally started with a linear model and followed by a nonlinear model to model the residual.In this paper,we hold the view that the hybrid sequence which is started with a nonlinear model and followed by a linear model has a more accurate prediction effect.The empirical results show that: 1.The combined model can integrate the information characteristics of the time series better;the combination model is more accurate in predicting the closing price of stock index and the statistical index is significantly better than the single prediction model;2.The hybrid sequence has an impact on the prediction results of the model.The prediction effect of the GRANN-ARIMA model is slightly better than that of the ARIMA-ANN model.