Board logo

标题: 请问二级NOTE 153页的一个问题 [打印本页]

作者: zmmengxue    时间: 2009-2-4 21:20     标题: 请问二级NOTE 153页的一个问题

各位高人,我对This makes sense because SEE measures the variability of the data about the regression line, and the more variable the data, the less confidence there is in the regressin model to edtimate a coefficent.这句话不是太理解,当the variability of the data 大时,SEE也就高,SEE高了,就意味着confidence interval widens. When confidence interval widens, why shouldnot there is more confidence in the regression model .(我想问的是,置信期间越大, 可信度是不是就越高)

原文来自于The standard error of the regression coefficient is denoted as sb1 . It is a function of the SEE: as SEE rises, sb1 also increases, and the confidence interval widens. This makes sense because SEE measures the variability of the data about the regression line, and the more variable the data, the less confidence there is in the regressin model to estimate a coefficent.(见NOTES 1 第153页)


作者: zyp444    时间: 2009-2-4 22:54

我觉得是不是可以这么来考虑:

confidence interval,即置信区间,单独来讲,它衡量的是样本统计值与总体参数值间误差范围,它越大则误差范围越大;而REGRESSION LINE是对总体参数值的一个描述,因此当the variability of the data 大时,SEE也就高,SEE高了,就意味着confidence interval widens,则误差范围越大,即是说通过REGRESSION LINE来做的估计就会有较大的误差,也就是书上所说的the less confidence there is in the regressin model to estimate a coefficent.

而至于说confidence interval越大,置信水平就越高,可以理解为对于事先给定的误差范围(即confidence interval),其越大,最后估计出来的变量值落在此区间的可能性就越高,因此可信度也就越大。

这是我的一点浅见,不知道正确与否,还请高手赐教啊。


作者: zmmengxue    时间: 2009-2-9 16:24

谢谢!
作者: Dong86    时间: 2009-2-10 13:39

你看错了吧。。。把less看成more了? 你再好好看看

This makes sense because SEE measures the variability of the data about the regression line, and the more variable the data, the less confidence there is in the regressin model to edtimate a coefficent.

这句话没问题啊。

SSE goes up --->more variability--->wider CI--->"LESS" confidence in the regression model.

没看到说wider CI--->more confidence哦





欢迎光临 FRM论坛 (http://frmspace.com/) Powered by Discuz! 7.2