[求助]做logit模型时，遇到的问题

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

在结果下面提示：“Note: 132 failures and 4352 successes completely determined.“不知道这里的失败和成功说明说明？？请指教，谢谢！

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0 回帖

关键词：logit模型 logit Log COMPLETELY Determined 模型 logit

. logit mode cost timehour

Iteration 0: log likelihood = -16941.18
Iteration 1: log likelihood = -9427.6488
Iteration 2: log likelihood = -6772.721
Iteration 3: log likelihood = -5353.1012
Iteration 4: log likelihood = -4677.5027
Iteration 5: log likelihood = -4397.6432
Iteration 6: log likelihood = -4324.1655
Iteration 7: log likelihood = -4317.6973
Iteration 8: log likelihood = -4317.6342

Logistic regression Number of obs = 31547
LR chi2(2) = 25247.09
Prob > chi2 = 0.0000
Log likelihood = -4317.6342 Pseudo R2 = 0.7451

------------------------------------------------------------------------------
mode | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
cost | -7.612456 .1288301 -59.09 0.000 -7.864959 -7.359954
timehour | 27.89635 .5787914 48.20 0.000 26.76194 29.03076
_cons | 1.753138 .0749419 23.39 0.000 1.606255 1.900022
------------------------------------------------------------------------------
Note: 132 failures and 4352 successes completely determined.

应该是数据的问题

你的检查数据

可能与解释变量中的某个值能够完全预测出model变量的值，所以倍剔除了。

谢谢版主！分别对时间和费用做logit模型发现费用成功的为0（Note: 1160 failures and 0 successes completely determined.），是否说明费用数据有问题？

logit mode timehour

Logistic regression Number of obs = 31547
LR chi2(1) = 823.64
Prob > chi2 = 0.0000
Log likelihood = -16529.362 Pseudo R2 = 0.0243

------------------------------------------------------------------------------
mode | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
timehour | 1.240174 .0474142 26.16 0.000 1.147244 1.333104
_cons | .5308182 .0280028 18.96 0.000 .4759336 .5857028
----------------------------------------------------------------------------

. logit mode cost

Logistic regression Number of obs = 31547
LR chi2(1) = 22737.22
Prob > chi2 = 0.0000
Log likelihood = -5572.5685 Pseudo R2 = 0.6711

------------------------------------------------------------------------------
mode | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
cost | -3.473881 .0473156 -73.42 0.000 -3.566618 -3.381144
_cons | 6.794507 .0740081 91.81 0.000 6.649454 6.939561
------------------------------------------------------------------------------
Note: 1160 failures and 0 successes completely determined.

要让变量尽量的变化，你可以考虑变量的重新分组