Logit模型
选择实验获得的数据主要通过离散选择模型来完成。离散选择模型中,最主要的是logit模型。
之前已经介绍了二项logit模型回归的STATA实现(有修改),多项logit模型详解,多项logit模型回归系数解读,多项logit模型回归的检验
继续认识混合logit模型。
. cmset id insurance
caseid variable: id
alternatives variable: insurance
. cmmixlogit choice, random(deductible premium) basealternative(5)
Fitting fixed parameter model:
Fitting full model:
Iteration 0: log simulated likelihood = -295.88154 (not concave)
Iteration 1: log simulated likelihood = -295.61382
Iteration 2: log simulated likelihood = -294.83963
Iteration 3: log simulated likelihood = -294.30391
Iteration 4: log simulated likelihood = -294.29896
Iteration 5: log simulated likelihood = -294.29896
Mixed logit choice model Number of obs = 1,250
Case ID variable: id Number of cases = 250
Alternatives variable: insurance Alts per case: min = 5
avg = 5.0
max = 5
Integration sequence: Hammersley
Integration points: 579 Wald chi2(2) = 67.93
Log simulated likelihood = -294.29896 Prob > chi2 = 0.0000
-------------------------------------------------------------------------------
choice | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
insurance |
deductible | -1.170465 .3603088 -3.25 0.001 -1.876657 -.4642729
premium | -2.884635 .3524249 -8.19 0.000 -3.575376 -2.193895
--------------+----------------------------------------------------------------
/Normal |
sd(deductible)| .8917091 .4088678 .3630186 2.19037
sd(premium)| .7575605 .3986796 .2700583 2.125089
--------------+----------------------------------------------------------------
Health |
_cons | 4.146376 .895823 4.63 0.000 2.390595 5.902156
--------------+----------------------------------------------------------------
HCorp |
_cons | 3.686473 .7823089 4.71 0.000 2.153176 5.21977
--------------+----------------------------------------------------------------
SickInc |
_cons | 2.813831 .6328887 4.45 0.000 1.573392 4.05427
--------------+----------------------------------------------------------------
MGroup |
_cons | 1.413957 .4315399 3.28 0.001 .5681547 2.25976
--------------+----------------------------------------------------------------
MoonHealth | (base alternative)
-------------------------------------------------------------------------------
LR test vs. fixed parameters: chi2(2) = 4.48 Prob > chi2 = 0.1064
Note: LR test is conservative and provided only for reference.
. di normal(-1.170465/0.8917091)
.09465744
. di normal(-2.884635/0.7575605)
.00007011
. cmmixlogit choice, random(deductible premium) casevars(income) basealternative
> (5)
Fitting fixed parameter model:
Fitting full model:
Iteration 0: log simulated likelihood = -290.37017 (not concave)
Iteration 1: log simulated likelihood = -290.35564
Iteration 2: log simulated likelihood = -289.03147
Iteration 3: log simulated likelihood = -288.91966
Iteration 4: log simulated likelihood = -288.91924
Iteration 5: log simulated likelihood = -288.91924
Mixed logit choice model Number of obs = 1,250
Case ID variable: id Number of cases = 250
Alternatives variable: insurance Alts per case: min = 5
avg = 5.0
max = 5
Integration sequence: Hammersley
Integration points: 579 Wald chi2(6) = 62.87
Log simulated likelihood = -288.91924 Prob > chi2 = 0.0000
-------------------------------------------------------------------------------
choice | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
insurance |
deductible | -1.155629 .3668827 -3.15 0.002 -1.874706 -.4365518
premium | -3.013088 .3897016 -7.73 0.000 -3.776889 -2.249287
--------------+----------------------------------------------------------------
/Normal |
sd(deductible)| .8472954 .4398034 .3063401 2.343505
sd(premium)| .8579541 .4192563 .3292362 2.235736
--------------+----------------------------------------------------------------
Health |
income | .6444183 .2753344 2.34 0.019 .1047728 1.184064
_cons | 1.237433 1.45387 0.85 0.395 -1.612099 4.086965
--------------+----------------------------------------------------------------
HCorp |
income | .4975011 .2453446 2.03 0.043 .0166346 .9783677
_cons | 1.483269 1.254495 1.18 0.237 -.9754962 3.942035
--------------+----------------------------------------------------------------
SickInc |
income | .1858664 .2281453 0.81 0.415 -.2612902 .633023
_cons | 2.093464 1.177981 1.78 0.076 -.2153365 4.402264
--------------+----------------------------------------------------------------
MGroup |
income | .1461937 .2188599 0.67 0.504 -.2827639 .5751512
_cons | .7965893 1.108163 0.72 0.472 -1.375371 2.968549
--------------+----------------------------------------------------------------
MoonHealth | (base alternative)
-------------------------------------------------------------------------------
LR test vs. fixed parameters: chi2(2) = 4.03 Prob > chi2 = 0.1336
Note: LR test is conservative and provided only for reference.
. di .8472954/.4398034
1.9265322
. di .8579541/.4192563
2.0463714
参考文献:
陈强主编,高级计量经济学及stata应用,高等教育出版社
Stata 官网
联系客服