时空加权回归模型的影响分析  

Influence analysis of geographically and temporally weighted regression model

作　　者：玄海燕[1] 李帅峰[2] 张运虎[1] 

机构地区：[1]兰州理工大学理学院,甘肃兰州730050 [2]海南大学应用科技学院(儋州校区),海南海口571737

出　　处：《兰州理工大学学报》2013年第5期135-138,共4页Journal of Lanzhou University of Technology

基　　金：国家自然科学基金(11261031)

摘　　要：基于数据删失模型,探讨各组观测数据对模型拟合的影响程度,给出Cook统计量;基于均值漂移模型,利用两步估计法对模型进行拟合,研究异常点检验问题,并构造检验统计量.The fitting influence degree of every grouped observation data on model fitting was explored based on the data deletion model and the Cook statistic was given. The problem of outlier test was explored by means of model fitting with two-step estimation method on the basis of mean shift model and the test statistic was constructed.

关 键 词：时空加权回归 数据删失模型 均值漂移模型 异常点检验 

分 类 号：O212.1[理学—概率论与数理统计]