MODEL SELECTION FOR EXTENDED QUASI-LIKELIHOOD MODELS IN SMALL SAMPLES
MODEL SELECTION FOR EXTENDED QUASI-LIKELIHOOD MODELS IN SMALL SAMPLES
复制标题
DOI：
10.2307/2533006
复制
发表时间：
1995-09-01
期刊：
影响因子：
1.9
通讯作者：
TSAI, CL
中科院分区：
文献类型：
--
作者：
HURVICH, CM;TSAI, CL
我们发展了一个小样本准则（AIC（c））用于扩展拟似然模型的选择。与Akaike信息准则（AIC）相比，AIC（c）为期望的Kullback-Leibler信息提供了一个更接近无偏的估计。因此，在小样本情况下，它通常选择比AIC更好的模型。对于Logistic回归模型，Monte Carlo结果表明AIC（c）优于AIC，Pregibon（1979，Data Analytic Methods for Generalized Linear Models.博士论文。多伦多大学）C*（p），以及Hosmer等人（1989，Biometrics 45，1265-1270）的C-p选择标准。两个例子。
We develop a small sample criterion (AIC(c)) for the selection of extended quasi-likelihood models. In contrast to the Akaike information criterion (AIC), AIC(c) provides a more nearly unbiased estimator for the expected Kullback-Leibler information. Consequently, it often selects better models than AIC in small samples. For the Logistic regression model, Monte Carlo results show that AIC(c) outperforms AIC, Pregibon's (1979, Data Analytic Methods for Generalized Linear Models. Ph.D, thesis. University of Toronto) C*(p), and the C-p selection criteria of Hosmer et al. (1989, Biometrics 45, 1265-1270). Two examples are presented.