GLIM AND NORMALIZING CONSTANT MODELS IN SPATIAL AND DIRECTIONAL-DATA ANALYSIS
GLIM AND NORMALIZING CONSTANT MODELS IN SPATIAL AND DIRECTIONAL-DATA ANALYSIS
复制标题
DOI：
10.1016/0167-9473(92)90140-b
复制
发表时间：
1992-04-01
影响因子：
1.8
通讯作者：
LAWSON, AB
中科院分区：
文献类型：
--
作者：
LAWSON, AB
本文提出了一种计算GLIM中归一化常数的方法。基于加权泊松模型和适当选择的虚拟权重，可以拟合各种各样的归一化常数模型。例子给出的LR测试“冯Misesness”相当于考克斯（1975年）的评分测试，估计的Kappa的Fisher分布，和参数估计的空间异质泊松过程。
A method is presented for evaluation of normalising constants in GLIM. Based on a weighted Poisson model and suitably chosen dummy weights a wide variety of normalising constant models can be fitted. Examples are given of an LR test for 'von Misesness' equivalent to Cox's (1975) score test, the estimation of kappa for a Fisher distribution, and parameter estimation for a spatial Heterogeneous Poisson Process.