Analysis、Modeling and Application of Structure-Transition Model of Time Series
|Course||Technology Economics and Management|
|Keywords||Threshold autoregressive model Threshold error correction model Smooth transition autoregressive model Smooth transition error correction model|
Financial analysis and modeling of time series is a very important area of ??research in financial econometrics. Nonlinear problems in the financial data and financial time series analysis of nonlinear econometric model is a new research topic in this field. Existing research nonlinear time series structure transformation model has two main streams, one in a time of structural change in the turning point of the study, followed by regression method is Tong threshold was first proposed in 1978. This method of the structural changes to the variable is the turning point. Time as a turning point in the analysis method identified the point in time of the change of the structure of subjective and less objective method, therefore, concluded there quite divided. In addition, when the model important explanatory variable in the short time persists more substantial changes, the traditional time-series model may not be able to diagnose the structural change of the model. In contrast to the methods of analysis of the variables as a turning point, can avoid the above-mentioned deletion. Variables can be directly observed to, so we can use directly. Detailed classification of the time series structure transformation model. And focused on the model of threshold cointegration relationship modeling, parameter estimation, conversion interval to determine the scope and inspection issues. The main work of this paper are as follows: 1. Door limit cointegration modeling subdivision. Use of the theoretical modeling makes it more selective. China in recent years, a continuous decrease of the nominal interest rate, and mild deflation in the economy which the nominal interest rate and the price level of the same downward trend, but this representation is not enough to judge the \exist. To this end, in this paper, the use of the threshold cointegration modeling system modeling method to establish the threshold vector error correction model, re-evaluate our long-term relationship between the nominal interest rate and inflation rate. Synchronization estimation algorithm for bivariate door 3.Hansen and Seo (hereinafter referred to as the HS) in 2002, the use of raster Find (grid search) limit vector error correction model and has a single cointegration vector cointegration vector and threshold . Use this method to deal with multi-variable vector error correction model large systems with multiple cointegration vector is generated on the difficulties in the calculation. This is because this algorithm involves a to-door limit cointegration vector while Find. The HS algorithm becomes difficult for large systems in terms of. For this reason, in this article the HS algorithm to improve in order to make this algorithm when processing large system flexibility.