AR模型仿真相关高斯序列的阶数选择

采用模型仿真相关高斯序列,具有灵活性强、效率高的优点,但如何选择合适的阶数一直是模型谱估计中的关键问题.比较了四种信息量准则和修正的Levison递推方法确定阶数的优缺点,并讨论了采样点数对阶数选择和仿真结果的影响.通过选择合适的信息量准则和采样点数来确定模型阶数,可以有效地仿真功率谱服从高斯谱、n次方谱和指数谱等多种谱形的相关高斯序列,大大降低了仿真过程的计算量.

Order Selection of Autoregressive Model in Simulation of Correlated Gauss Sequences

Simulation of correlated Gaussian sequences by autoregressive model has the merits of flexibility and high efficiency,but how to select a suitable order is a pivotal problem of estimating autoregressive model power spectrum. The merits and weaknesses of four information criteria and the modified Levison algorithm are compared in this paper,and the influence of samples for order selection and simulation result are discussed.Through selecting order using suitable information criteria and samples,we have simul...