The Applicability of Time Series Models on Shanghai Securities Composite Index Short-term Forecasting

摘要

Efforts are made in this paper to explore a short-term forecast (04/23/2012 to 05/04/2012) of Shanghai Securities Composite Index (SSCI) closing price. According to the theory of time series, taking seasonal cycle, heteroscedasticity and leverage effect into account, ARIMA product season model, GARCH model and GJR model are built gradually for forecasting by using the data of SSCI closing price between 01/04/2010 and 04/20/2012 as a training sample. GJR(2,1,1) Model is the chosen one for its highest relative nearness degree and best forecasting performance. It points out the significant effects of leverage effect and policy factors on SSCI, and the necessity for improving the system, establishing the short mechanism and reducing the government intervention.