怎么用STATA做加权最小二乘法来降低异方差

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

怎么用STATA做加权最小二乘法来降低异方差

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏8 回帖

关键词：加权最小二乘法 最小二乘法 Stata 最小二乘 tata

 你去查查wls0命令，这个的单独下载

兄弟能不能说清楚点啊

不明白你的意思

wls,没有这个命令哦

reg y x1 x2 [aw=v], where v is the weight!

----------------------------------------------------------------------------------------------------------
help for wls0
----------------------------------------------------------------------------------------------------------

Weighted least squares regression
---------------------------------

 wls0 dvar ivars [if] [in], wvar(varlist) type(wlstype)
     [ noconst robust hc2 hc3 graph ]

Description
-----------

wls0 computes various weighted least squares regressions. 
     The command generates two variables _wls_wgt and
     _wls_res, the wls weights and residuals respectively.

Options
-------

wvar(varlist) variables in the weightling equation.

type(wlstype) the wls type.  The choices are:
    abse  - absolute value of residual
    absen - absolute value of residual, ala Neter et al
    e2    - residual squared
    loge2 - log residual squared
    xb2   - fitted value squared

noconst do not include a constant in the weighting equation.

robust using robust standard errors in the final regression.

graph plots the wls residuals vs fitted.

Examples
--------

 . wls0 exp age ownrent income incomesq, wvar(income incomesq) type(abse)
 . wls0 exp age ownrent income incomesq, wvar(income incomesq) type(abse) noconst

Author
------

      Philip B. Ender
      UCLA Department of Education
      UCLA Academic Technology Services
      ender@ucla.edu

g id=_n

xtgls y x1-xn,i(id) p(h)

*可能达到目的（其中y是因变量，x1……xn是自变量）。 

wlso在那里下载啊

（2）纠正异方差的常用套路（构造h值）
reg y x1 x2 x3
predict u,resid
gen usq=u^2
gen logusq=log(usq)
reg logusq x1 x2 x3
predict  g
gen h=exp(g)
reg y x1 x2 x3 [aw=1/h]
异方差hausman检验：
reg y x1 x2 x3
est store A（将上述回归结果储存到A中）
reg y x1 x2 x3 [aw=1/h]
est store B
hausman A B

seems make sense. which one is  reasonable?

vwls  y x
与
y x [aw=weight]
是等价的吗？