请教Jarque-Bera Probability命令

相似文件 换一批

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

让高手见笑了，初学者请问一下，STATA里J-B test 得出J-B 值和 probability的命令语句是什么呢。在网上找了很久，都没有相关的内容。

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5 回帖

关键词：Probability bability Ability Jarque Bera 初学者 网上

------------------------------------------------------------------------------------
help for jb6
------------------------------------------------------------------------------------

Jarque-Bera test for normality
-----------------------------------------------

        jb6 varname


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

jb6 calculates the Jarque-Bera asymptotic test for normality on the     specified variable in level form.  

Options
-------

No options available

Example
-------
. jb6 varname


References
----------
  See C.M. Jarque and A.K. Bera.  1987.  "A Test for Normality of   Observations and Regression Residuals.  International Statistical   Review 55:163-172; In Damodar Gujarati.  1995.    Basic Econometrics.  p. 143.

  jb6 draws on the procedure originally written  by Gregorio Impavido,  gimpavido@worldbank.org
        The formula and code were slightly   modified by J. Sky David, jsdavid@polisci.tamu.edu

food.rar (580 Bytes) 本附件包括：

• food.dta

2012-7-22 10:20:25 上传

* open data
use food, clear

* log of income
gen lincome = ln(income)

* linear-log regression
reg food_exp lincome
predict lyhat
predict lehat, resid

* slope = beta2/x
summarize income
return list
scalar xbar = r(mean)
lincom lincome/xbar
lincom lincome/10
lincom lincome/20
lincom lincome/30

* fitted value plot
twoway (scatter food_exp income, sort)                                 ///
       (line lyhat income, sort lwidth(medthick)),         ///
            xtitle(Income) ytitle(Food Expenditure) ylabel(0(100)600) ///
                title(Linear Log Model)
graph save linlog, replace

* linear relationship
quietly reg food_exp income
predict yhat
predict ehat, resid

* linear and linear-log fitted lines
twoway (scatter food_exp income, sort)                                 ///
(line lyhat income, sort lwidth(medthick))                         ///
(line yhat income, sort lpattern(dash) lwidth(medthick)), ///
xtitle(Income) ytitle(Food Expenditure) ylabel(0(100)600) ///
title(Linear Log Model)
graph save linlog2, replace

* plot linear-log model residuals               
twoway (scatter lehat income, sort) ,         ///
       xtitle(Income) ytitle(Residuals) ///
           title(Linear Log Model Residuals)
graph save linlog_residual, replace

* analyze residuals from original equation
histogram ehat, percent title(Linear Model Residuals)
graph save olsehat_hist, replace

* Jarque-Bera test of error normality
summarize ehat, detail
return list

scalar jb = (r(N)/6)*( (r(skewness)^2) + ((r(kurtosis)-3)^2)/4 )
di "Jarque-Bera Statistic = " jb
scalar chic = invchi2tail(2,.05)
di "Chi-square(2) 95th percentile = " chic
scalar pvalue = chi2tail(2,jb)
di "Jarque-Bera p-value = " pvalue

good

蓝色 发表于 2012-7-22 10:15
------------------------------------------------------------------------------------
help for jb6
...

请问jb6的结果应该怎么看呢？jb6的output标明的不是很清楚。。。谢谢！

使用藤椅楼层大师的如下代码，跑成功啦

sum diff, detail
scalar jb = (r(N)/6)*( (r(skewness)^2) + ((r(kurtosis)-3)^2)/4 )
di "Jarque-Bera Statistic = " jb
scalar chic = invchi2tail(2,.05)
di "Chi-square(2) 95th percentile = " chic
scalar pvalue = chi2tail(2,jb)
di "Jarque-Bera p-value = " pvalue

  命令格式：jb6 varname