一类非正态误差下的AR模型定阶研究

Research on AR Model Order Determination in a Class of Non-normal Errors

对一类非正态误差的AR模型，在待定阶数P的情况下，给出误差项中未知实函数依概率有界的定理，可把非正态误差转化为正态情况。最后运用正态误差下AR模型的方法确定阶数和参数，并给出一个算例。

For the AR model of a class of non-normal errors, in the case of p-order to be determined, we propose a theorem that the unknown function in errors is probability bounded. Under certain conditions the non-normal error can be turned into a normal situation. Finally, the AR model in normal errors is used to determine the order and parameter, and an example is given.