基于ARIMA模型的春节因素调整方法研究

目的 研究基于ARIMA模型的春节因素调整方法.方法 构建通用的春节因素变量,将其作为回归变量纳入季节性ARIMA回归模型(regARIMA或TRAMO),采用AIC或BIC对模型的效果进行判断,确定最优模型.采用广义最小二乘法或最大似然法进行参数估计,并根据估计出的回归系数计算春节因素的影响程度.通过实例分析对上述方法进行实证.结果 实例分析表明,引入春节因素变量后的季节调整方法能有效地消除春节因素对时间序列的影响.并能定量分析春节因素的影响程度.结论 构建的春节因素变量具有较好的适用性,基于ARIMA模型的春节因素调整方法能有效地运用于时间序列的季节调整,为分析春节因素的影响提供了一种新的方法.

Study on the Methods of the Chinese New Year Factor's Adjustment Based on the ARIMA Models

Objective To study the methods of the Chinese New Year (CNY) Factor's Adjustment based on the ARIMA models. Methods First, a common regressor for CNY was created. Then, the re-gressor was included in the seasonal ARIMA regressive model(regARIMA or TRAMO) ,AIC or BIC was used for model selection,and the generalized least squares method or maximum likelihood method was used for the earl-mation of model parameter. The estimated regressive coefficient was used for analyzing the degree of the CNY factor. A case was analyzed with the adjustment methods. Results The analysis on the case showed that the methods of the CNY factor's adjustment could remove the effects of the CNY factor on the time series, and the degree of the effects could be esti-mated in quantity. Conclusion The regressor for CNY is applicable,and the methods of the CNY factor's adjustment based on the ARIMA models can be used in seasonal adjustment on the time series. It's a new approach to analyze the effects of the CNY factor.