论文检索

自回归模型阶数估计：max x 2 方法AI 解读

We develop a method of choosing the order of an autoregression to be used in conjunction with the subset autoregression technique (McClave 1975). A simulation is performed to compare this method (the max x 2 method) with Akaike's (1969, 1970) FPE technique of estimating order. The max x 2 outperforms FPE when used in conjunction with subset autoregression.

摘要暂缺

摘要暂缺

用自回归模型近似移动平均过程的偏差AI 解读

A rather extensive empirical study of autoregressive approximations to first-order moving averages, and in particular their use in estimating the moving average coefficient, shows a significant bias to be present. This is of special importance now that autoregressive spectral estimators are becoming popular. Several conjectures are made about the cause of bias, and some procedures are developed to reduce this bias.