摘要
研究一般线性结构EV模型。给出了线性EV模型与线性回归模型的区别,根据最小二乘方法给出模型的参数估计,并且对结构EV模型进行了统计诊断,证明了基于数据删除模型与均值漂移模型的参数估计的等价性,通过均值漂移模型分析其异常点。通过实例验证以上所得结论。
The article mainly studies the general linear structure EV model. First of all, it analyzes the difference between the linear EV models and linear regression model, and then, based on the least square method, the parameter estimation of EV model is given. Meanwhile, the EV model is statisti- cally diagnosed and the equivalence of the estimated parameters between data deletion model and mean shift model is improved. By F-test statistics, the breakdown points of the EV model are found out. Finally, through instance, the above conclusion is validated.
出处
《重庆理工大学学报（自然科学）》
CAS
2012年第5期101-104,共4页
Journal of Chongqing University of Technology：Natural Science
基金
国家自然科学基金资助项目(50979029)
关键词
线性EV模型
参数估计
均值漂移模型
异常点
linear EV model
parameter estimation
mean shift model
breakdown points
