JOINT ESTIMATION OF MODEL PARAMETERS AND OUTLIER EFFECTS IN TIME-SERIES
JOINT ESTIMATION OF MODEL PARAMETERS AND OUTLIER EFFECTS IN TIME-SERIES
复制标题
DOI：
10.2307/2290724
复制
发表时间：
1993-03-01
影响因子：
3.7
通讯作者：
LIU, LM
中科院分区：
文献类型：
--
作者：
CHEN, C;LIU, LM
时间序列数据通常会受到不受控制或意外的干预，从而产生各种类型的异常观察结果。时间序列中的异常值，根据其性质，可能会对时间序列分析标准方法在模型识别、估计和预测方面的有效性产生中等到显着的影响。在本文中，我们使用迭代异常值检测和调整程序来获得模型参数和异常值效应的联合估计。考虑了四种类型的异常值，并讨论了杂散和掩蔽效应问题。该程序与早期文献中提出的程序之间的主要区别包括（a）异常值的类型和影响是基于模型参数污染较少的估计而获得的，（b）使用多元回归同时估计异常值影响，以及（c）联合估计模型参数和异常值影响。通过模拟研究研究从小到大样本量情况下的检验统计量的抽样行为。该程序的性能通过一组有代表性的异常情况进行检查。我们发现所提出的程序在检测出口和获得无偏参数估计方面表现良好。使用一个例子来说明所提出的程序的应用。事实证明，该过程可以有效避免虚假异常值和掩盖效应。从所提出的程序获得的模型参数估计通常非常接近使用干预模型合并异常值的精确最大似然法估计的参数。
Time series data are often subject to uncontrolled or unexpected interventions, from which various types of outlying observations are produced. Outliers in time series, depending on their nature, may have a moderate to significant impact on the effectiveness of the standard methodology for time series analysis with respect to model identification, estimation, and forecasting. In this article we use an iterative outlier detection and adjustment procedure to obtain joint estimates of model parameters and outlier effects. Four types of outliers are considered, and the issues of spurious and masking effects are discussed. The major differences between this procedure and those proposed in earlier literature include (a) the types and effects of outliers are obtained based on less contaminated estimates of model parameters, (b) the outlier effects are estimated simultaneously using multiple regression, and (c) the model parameters and the outlier effects are estimated jointly. The sampling behavior of the test statistics for cases of small to large sample sizes is investigated through a simulation study. The performance of the procedure is examined over a representative set of outlier cases. We find that the proposed procedure performs well in terms of detecting outlets and obtaining unbiased parameter estimates. An example is used to illustrate the application of the proposed procedure. It is demonstrated that this procedure performs effectively in avoiding spurious outliers and masking effects. The model parameter estimates obtained from the proposed procedure are typically very close to those estimated by the exact maximum likelihood method using an intervention model to incorporate the outliers.