Research And Application Of Time Series Analysis In Stock

The development of the stock market can reflect the country’s economic strength, thecompany’s development potential, people’s income level and so on. So the research ofstock market attracts many scholars’ attention. But there is a greater difficulty in predictingstock market’s movement accurately. Establishing an accurate prediction model hasbecome the primary task of financial experts and investors. At present, there are manypredicting methods on stock price movement, but their fitting effect is poor, and theprediction accuracy is low. On account of the time series analysis model has wideapplication range, small limitation and high accuracy on short-term prediction, time seriesanalysis model has become one of the most popular prediction method in financialprediction. In this paper, ARIMA model based on wavelet analysis and metabolicGM(1,1)-TARMA model are established, and the Shanghai composite index is used to testthe prediction effect. The main contents are as follows:1) The movement of Shanghai composite index is analyzed deeply. Because of thevolatility and non-stationary of Shanghai composite index, ARIMA model based onwavelet analysis is established to fit the monthly average closing price of Shanghaicomposite index from December1990to March2013, and the simulation results show thatthe model has higher prediction precision and stronger practicability. The waveletdecomposition can solve many problems effectively, such as the original sequencebecoming white noise sequence after difference, and low frequency sequence becomingmore smooth which can better reflect the linear relationship of the sample data. It isbeneficial to set up the time series model.2) Due to the stock price index system is a system that the quantity of sample data isrelatively large, and the grey prediction model is mainly used in small sample system. Inorder to applying grey prediction model to larger data sample system, GM (1,1)-TARMAhybrid model is established using the metabolism GM (1,1) grey model and the TARMA model to fitting Shanghai composite index. This paper introduced the metabolism GM (1,1)model, and use the TARMA models to modify the residual of metabolic GM (1,1) model tofurther improve the prediction accuracy. Finally, the average monthly closing price ofShanghai composite index from June2013to August2013are predicted using this modelusing the Matlab and Eviews. Result show that, the GM (1,1)-TARMA hybrid model canpredict the trend of the stock price index better.