逆自相关函数及其在ARMA模型识别中的作用

Improve the Effectiveness of Identifying ARMA Models with the Application of the Inverse Autocorrelations

目的 介绍逆自相关函数的意义 ,说明其在ARMA模型识别中的作用。方法 分析逆自相关函数以了解时序的结构特征 ,获得简约的拟合模型形式。结果 实例分析表明 ,逆自相关函数在疏系数模型的识别中 ,明确地补充刻划了模型结构 ,有可靠的应用价值。结论 在时间序列的ARMA模型识别中 ,考虑自相关函数与偏自相关函数的同时 ,充分利用逆自相关函数提供的信息 ,有助于选得较优模型。

Objective The concept of the inverse autocorrelations was given,and its applications in indentifying ARMA models were displayed.Methods As to get a parsimonious ARMA model for a time series,some characteristics of the inverse autocorrelations were discussed in detail.Results As is shown in the given example,the inverse autocorrelations are of great value in identifying a desirable ARMA model,especially those with sparse parameters.Conclusion It's necessary to consider the inverse autocorrelations besides the autocorrelations and the partial auto correlations when identifying ARMA models.