Я запустил отрицательную биномиальную модель в R, используя следующий код.
myModel <- glm.nb(V7 ~V11 + V8 + V3 + V10 + V12 + V5, data = data)
summary(myModel)
V7 является зависимой переменной. Другие переменные являются независимыми переменными.
Я получил следующий вывод. Но вывод R отличается от вывода Stata. Что я должен сделать, чтобы исправить этот код?
myModel <- glm.nb(V7 ~ V11 + V8 + V3 + V10 + V12 + V5, data = data)
> summary(myModel)
Call:
glm.nb(formula = V7 ~ V11 + V8 + V3 + V10 + V12 + V5, data = data,
init.theta = 0.4931977401, link = log)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.3363 -1.0487 -0.7826 0.1631 2.4126
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -7.046e-01 3.824e-01 -1.842 0.065408 .
V11 -1.705e-01 2.909e-01 -0.586 0.557875
V8 -1.549e-06 4.371e-06 -0.354 0.723149
V3 4.525e-02 1.168e-02 3.873 0.000108 ***
V10 3.050e-04 3.275e-03 0.093 0.925802
V12 -1.965e-03 6.121e-03 -0.321 0.748157
V5 5.378e-03 5.954e-03 0.903 0.366423
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for Negative Binomial(0.4932) family taken to be 1)
Null deviance: 270.87 on 285 degrees of freedom
Residual deviance: 233.16 on 279 degrees of freedom
AIC: 726.49
Number of Fisher Scoring iterations: 1
Theta: 0.4932
Std. Err.: 0.0922
2 x log-likelihood: -710.4890
Stata output
nbreg V7 V11 V8 V3 V10 V12 V5
Fitting Poisson model:
Iteration 0: log likelihood = -838.04152
Iteration 1: log likelihood = -666.38251
Iteration 2: log likelihood = -432.95361
Iteration 3: log likelihood = -422.79171
Iteration 4: log likelihood = -422.19242
Iteration 5: log likelihood = -422.18128
Iteration 6: log likelihood = -422.18126
Fitting constant-only model:
Iteration 0: log likelihood = -390.16992
Iteration 1: log likelihood = -380.55285
Iteration 2: log likelihood = -371.83158
Iteration 3: log likelihood = -371.83069
Iteration 4: log likelihood = -371.83069
Fitting full model:
Iteration 0: log likelihood = -359.49706
Iteration 1: log likelihood = -355.85223
Iteration 2: log likelihood = -355.24795
Iteration 3: log likelihood = -355.24464
Iteration 4: log likelihood = -355.24464
Negative binomial regression Number of obs = 286
LR chi2(6) = 33.17
Dispersion = mean Prob > chi2 = 0.0000
Log likelihood = -355.24464 Pseudo R2 = 0.0446
-----------------------------------------------------------------------------------
After_crashes | Coef. Std. Err. z P>|z| [95% Conf. Interval]
------------------+----------------------------------------------------------------
V11 | -1.55e-06 4.84e-06 -0.32 0.749 -.000011 7.95e-06
V8 | .0452497 .0124943 3.62 0.000 .0207613 .069738
V3 | .000305 .0030724 0.10 0.921 -.0057168 .0063268
V10 | -.0019654 .0058801 -0.33 0.738 -.0134903 .0095595
V12 | -.1704632 .2747938 -0.62 0.535 -.7090493 .3681228
V5 | .0053776 .0074238 0.72 0.469 -.0091729 .0199281
_cons | -.7046201 .3960827 -1.78 0.075 -1.480928 .0716878
------------------+----------------------------------------------------------------
/lnalpha | .7068451 .1878484 .3386689 1.075021
------------------+----------------------------------------------------------------
alpha | 2.027584 .3808785 1.403079 2.930055
-----------------------------------------------------------------------------------
LR test of alpha=0: chibar2(01) = 133.87 Prob >= chibar2 = 0.000