摘要
利用贝叶斯因子对AR模型定阶问题进行讨论,并给出了AR模型的贝叶斯因子的具体表达式,且用随机模拟对其与AIC(k),BIC(k)两种定阶准则进行比较得出,用贝叶斯因子方法的判别效果优于AIC(k)准则.
This paper takes advantage of the Bayesian factor to discuss the problem of selecting the order of AR models. The specific expression for the Bayesian factor of AR models is given and the simulation comparisons among Bayesian factor method, AIC（k） and BIC（k） are made. The results show that Bayesian factor method is better than the AIC（k）.
出处
《湖北工业大学学报》
2007年第1期13-15,共3页
Journal of Hubei University of Technology
