The Application Research Of Time Series Modeling Methods

In this paper, we research several application of stochastic trend and determine trend those two types time series from some ideas of traditional time series analysis-ARIMA model systems and grey time series analysis theories in order to achieve two objectives: first, recognize the date generation mechanism of observed sequence; second, forecast the possible value of the sequence in future.For stochastic trend time series,this paper from the idea of traditional time series analysis methods combine with nonparametric statistical methods to establish statistical model in order to research data generation mechanism of the sequence. Here, we take a commercial bank's management of cashnd as a example. Using a nonparametric software Eviews 6.0 to fit out the approximation expression form for net positions.s density function, to get the specific form of density function, to make use of the method of moment estimation for the parameters, get the estimated values of parameters in density function,so that we can obtain the probability density function of the net positions, then make the statistical inference, establish the model of cash generation so as to provided the scientific basis and methods for cash management.For determine trend time series, this paper from the idea of grey system theory research several sequences associate with cash other. First of all, accumulate(subtract)those time series one or more times,then construct new sequences according to the nature of the data and establish multiple linear(or nonlinear)regression models(differ from GM(1,1),GM(r,h)models of grey system theory)for all those sequence.Given the level of significance?, then carry out significance test for every regression equations in confidence level 1- ?.If test through, using differential and difference to establish single differential equation model then compose a system of differential equations model so as to reveal the dynamic relationships among all the time series and forecast the possible value in future. Finally, take GDP and visitors of Hainan province in 1995-2014 as a example to establish dualistic equations model and forecast the future.