The Applications Of Approximate Unbiased Estimation Of Outliers In Economic Time Series Models

Economic time series are susceptible to interference from various factors to produce different types of outliers,and their existence will have a greater impact on model identification and parameter estimation.The deviation of the estimation of the interference amplitude of the outlier will greatly affect the overall effect of the model.Therefore,the research on the interference amplitude of the outlier has great significance.Based on the minimum phase and reversible conditions of the autoregressive moving average model(ARMA),the self-regressive model(AR)is transformed by the squared term property of the autoregressive conditional heteroscedasticity model(ARCH)and the generalized autoregressive conditional heteroscedasticity model(GARCH).Through the least squares estimation of the additional outlier(AO)and the innovative outlier(IO)interference amplitude in the AR model and the correction of its deviation,it is generalized to obtain the interference amplitude of approximate unbiased estimation of the ARCH model and the outlier of the GARCH model.Finally,the data simulation and case demonstration show that the better the sample size is,the better the least squares estimation and the approximate unbiased estimation are.And the correction effect of the approximate unbiased estimation is always better than the least squares estimation.As the amplitude of the interference increases,the correction effect of the approximate unbiased estimation becomes more obvious.