NO.PZ2018101001000067
问题如下:
If the income of Paul’s company shows nonstationary and also has a serial correlation, which of the following models could be used to predict the income of his company?
选项:
A.Linear trend model.
B.AR(1) model.
C.First-differenced AR(2) model.
解释:
C is correct.
考点: AR模型假设及修正
解析:当我们发现一个时间序列呈现序列相关时,我们要用AR(2)模型进行修正,若这个序列不是协方差平稳的,我们需要对其进行一阶差分来修正这个模型。所以在以上三个选项中,C选项First-differenced AR(2) model是我们最有可能用到的模型。
老师好,我刚刚翻了下关于这道题之前的问题和回答,有老师说AR(1)和AR(2)没有太大的区别。因为题目说non-stationary,所以选一阶差分,所以就选了C。
但是不是如果选项中有“一阶差分AR(1)”、“一阶差分AR(3)”..也正确呢?谢谢