Prediction of the Stock Price of Shanghai Securities Composite Index by Using Time-series Model

Abstract

Investors collect information from the trading market and make investment decisions based on the collected information, i.e. belief in the future trend of securities prices. Therefore, some time series models are analyzed and methodology came into being and gradually developed. However, accurate trend prediction has long been a difficult problem. To improve the prediction accuracy, we must take advantage of the mathematical model (Time series model) to predict the stock price. When it comes to the time series model, ARIMA is impossible to be ignored. This paper has used the time series model including ARMA and ARIMA for predicting the stock price and it also includes parameter assignment about p、q、d for evaluating AIC and BIC by using ADF test、ACF、PACF. In addition, models are evaluated through experiments on real data sets composed of the true value of the Shanghai Securities Composite Index and fitting value of Shanghai Securities Composite Index and the table shows that the model we establish can minimize the error to a large degree. Therefore, the prediction result is in precision.