悬赏 180 个论坛币 未解决
想请问下,怎么在stata中用boxcox将一列本不是正态分布的数据转化为正态分布的?
做了命令如下:“boxcox absx2”(absx2为变量值),得到的数据结果如下:
. boxcox absx2
Fitting comparison model
Iteration 0: log likelihood = -939.70641
Iteration 1: log likelihood = -146.16592
Iteration 2: log likelihood = -128.71365
Iteration 3: log likelihood = -128.34386
Iteration 4: log likelihood = -128.34366
Iteration 5: log likelihood = -128.34366
Fitting full model
Iteration 0: log likelihood = -939.70641
Iteration 1: log likelihood = -146.16592
Iteration 2: log likelihood = -128.71365
Iteration 3: log likelihood = -128.34386
Iteration 4: log likelihood = -128.34366
Iteration 5: log likelihood = -128.34366
Number of obs = 204
LR chi2(0) = 0.00
Log likelihood = -128.34366 Prob > chi2 = .
------------------------------------------------------------------------------
absx2 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
/theta | -.0298817 .0259776 -1.15 0.250 -.0807969 .0210335
------------------------------------------------------------------------------
Estimates of scale-variant parameters
----------------------------
| Coef.
-------------+--------------
Notrans |
_cons | -1.316571
-------------+--------------
/sigma | 1.650568
----------------------------
---------------------------------------------------------
Test Restricted LR statistic P-value
H0: log likelihood chi2 Prob > chi2
---------------------------------------------------------
theta = -1 -614.50175 972.32 0.000
theta = 0 -129.01617 1.35 0.246
theta = 1 -939.70641 1622.73 0.000
---------------------------------------------------------