我试图用4个方程（2个直接影响和2个中介影响）估计一个看似无关的回归模型。我的模型中有3个虚拟变量，它们在输出中得到NA分数。虚拟变量来自测量购物渠道偏好的变量：1 =在线2 =离线3 =无偏好。首先，当我制作虚拟人物时，我仅制作了2个虚拟人物：Online_preference和Offline_preference，其中1代表Online / Offline偏好，其他代表零。结果对我来说似乎很奇怪，然后我添加了第三个哑元，即No_preference，表示无偏好为1，所有其他指示为0。
我还尝试将原始变量channel.preference包含为包含值1、2和3的因数。但是，截距变为1。因此没有要解释的截距。我的硕士论文需要这个。
库（系统调整）
eq1 <- loyalonline ~ Channel_preference_NO + Channel_preference_offline + Channel_preference_online + X.shop.on + X.shop.off + Integration + servqual_online + servqual_offline + assortment_online + assortment_offline + price_online + price_offline + image_offline + image_online
eq2 <- loyaloffline ~ Channel_preference_NO + Channel_preference_offline + Channel_preference_online + X.shop.on + X.shop.off + Integration + servqual_online + servqual_offline + assortment_online + assortment_offline + price_online + price_offline + image_offline + image_online
eq3 <- image_offline ~ Channel_preference_NO + Channel_preference_offline + Channel_preference_online + X.shop.on + X.shop.off + Integration + servqual_online + servqual_offline + assortment_online + assortment_offline + price_online + price_offline + image_online
eq4 <- image_online ~ Channel_preference_NO + Channel_preference_offline + Channel_preference_online + X.shop.on + X.shop.off + Integration + servqual_online + servqual_offline + assortment_online + assortment_offline + price_online + price_offline + image_offline
fitsur <- systemfit(list(eq1, eq2, eq3, eq4), "SUR", data=surveydata2)
summary(fitsur)
这为第一个方程式产生了以下输出：
SUR estimates for 'eq1' (equation 1)
Model Formula: loyalonline ~ Channel_preference_NO + Channel_preference_offline +
Channel_preference_online + X.shop.on + X.shop.off + Integration +
servqual_online + servqual_offline + assortment_online +
assortment_offline + price_online + price_offline + image_offline +
image_online
Estimate Std. Error t value Pr(>|t|)
(Intercept) -8.197137062 NA NA NA
Channel_preference_NO 8.492880678 NA NA NA
Channel_preference_offline 8.931039829 NA NA NA
Channel_preference_online 8.740189838 NA NA NA
X.shop.on -0.000937155 0.086172689 -0.01088 0.991334
X.shop.off 0.116146382 0.087283445 1.33068 0.184881
Integration -0.086697750 0.139339296 -0.62221 0.534548
servqual_online 0.407035659 0.162080604 2.51132 0.012857 *
servqual_offline -0.198559780 0.127637103 -1.55566 0.121445
assortment_online 0.262651479 0.123362918 2.12910 0.034527 *
assortment_offline 0.079556741 0.121200399 0.65641 0.512353
price_online 0.071331977 0.114409532 0.62348 0.533713
price_offline -0.001153320 0.110379210 -0.01045 0.991674
image_offline 0.153717670 0.135150835 1.13738 0.256805
image_online 0.285603534 0.143154315 1.99507 0.047456 *
As can be seen the first 3 variables are NA.
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.92452 on 191 degrees of freedom
Number of observations: 206 Degrees of Freedom: 191
SSR: 163.254828 MSE: 0.854737 Root MSE: 0.92452
Multiple R-Squared: 0.293031 Adjusted R-Squared: 0.241212
SUR estimates for 'eq2' (equation 2)
Model Formula: loyaloffline ~ Channel_preference_NO + Channel_preference_offline +
Channel_preference_online + X.shop.on + X.shop.off + Integration +
servqual_online + servqual_offline + assortment_online +
assortment_offline + price_online + price_offline + image_offline +
image_online
Estimate Std. Error t value Pr(>|t|)
(Intercept) 11.3664256 NA NA NA
Channel_preference_NO -11.3321726 NA NA NA
Channel_preference_offline -11.3754037 NA NA NA
Channel_preference_online -10.9301111 NA NA NA
X.shop.on 0.1025450 0.0701521 1.46175 0.145452
X.shop.off -0.0933838 0.0710563 -1.31422 0.190348
Integration 0.0336123 0.1134343 0.29631 0.767312
servqual_online -0.1408759 0.1319477 -1.06766 0.287019
servqual_offline 0.1672898 0.1039077 1.60998 0.109053
assortment_online 0.2332130 0.1004282 2.32219 0.021277 *
assortment_offline 0.1754046 0.0986677 1.77773 0.077039 .
price_online -0.0202389 0.0931393 -0.21730 0.828209
price_offline 0.1893764 0.0898583 2.10750 0.036379 *
image_offline 0.6943886 0.1100246 6.31122 1.8889e-09 ***
image_online -0.2262248 0.1165401 -1.94118 0.053709 .
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.75264 on 191 degrees of freedom
Number of observations: 206 Degrees of Freedom: 191
SSR: 108.195146 MSE: 0.566467 Root MSE: 0.75264
Multiple R-Squared: 0.462038 Adjusted R-Squared: 0.422607
SUR estimates for 'eq3' (equation 3)
Model Formula: image_offline ~ Channel_preference_NO + Channel_preference_offline +
Channel_preference_online + X.shop.on + X.shop.off + Integration +
servqual_online + servqual_offline + assortment_online +
assortment_offline + price_online + price_offline + image_online
Estimate Std. Error t value Pr(>|t|)
(Intercept) -43.5016667 NA NA NA
Channel_preference_NO 43.5551857 NA NA NA
Channel_preference_offline 43.6567223 NA NA NA
Channel_preference_online 43.5747272 NA NA NA
X.shop.on 0.0888251 0.0455753 1.94897 0.0527554 .
X.shop.off -0.0393647 0.0462627 -0.85090 0.3958874
Integration -0.1678208 0.0736773 -2.27778 0.0238413 *
servqual_online -0.0832320 0.0861682 -0.96592 0.3352971
servqual_offline 0.1905985 0.0653210 2.91788 0.0039448 **
assortment_online -0.2704448 0.0641071 -4.21864 3.7859e-05 ***
assortment_offline 0.3843459 0.0589268 6.52243 5.9783e-10 ***
price_online 0.0313554 0.0610573 0.51354 0.6081640
price_offline 0.0298582 0.0588896 0.50702 0.6127223
image_online 0.7071729 0.0659574 10.72165 < 2.22e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.516626 on 192 degrees of freedom
Number of observations: 206 Degrees of Freedom: 192
SSR: 51.245306 MSE: 0.266903 Root MSE: 0.516626
Multiple R-Squared: 0.419614 Adjusted R-Squared: 0.380317
SUR estimates for 'eq4' (equation 4)
Model Formula: image_online ~ Channel_preference_NO + Channel_preference_offline +
Channel_preference_online + X.shop.on + X.shop.off + Integration +
servqual_online + servqual_offline + assortment_online +
assortment_offline + price_online + price_offline + image_offline
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.56961e+01 4.43470e+05 0.00004 0.9999718
Channel_preference_NO -1.53648e+01 4.43470e+05 -0.00003 0.9999724
Channel_preference_offline -1.53712e+01 4.43470e+05 -0.00003 0.9999724
Channel_preference_online -1.55130e+01 4.43470e+05 -0.00003 0.9999721
X.shop.on -5.76136e-02 4.33280e-02 -1.32971 0.1851924
X.shop.off -3.14545e-02 4.37076e-02 -0.71966 0.4726109
Integration 1.56339e-01 6.95914e-02 2.24653 0.0258089 *
servqual_online 2.29307e-01 7.93302e-02 2.89054 0.0042887 **
servqual_offline 3.65699e-03 6.34783e-02 0.05761 0.9541191
assortment_online 3.26501e-01 5.80803e-02 5.62154 6.5952e-08 ***
assortment_offline -2.67714e-01 5.92847e-02 -4.51574 1.0996e-05 ***
price_online -2.42917e-02 5.76610e-02 -0.42128 0.6740188
price_offline -1.15770e-02 5.56388e-02 -0.20807 0.8353910
image_offline 6.30310e-01 5.87885e-02 10.72165 < 2.22e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.487743 on 192 degrees of freedom
Number of observations: 206 Degrees of Freedom: 192
SSR: 45.675433 MSE: 0.237893 Root MSE: 0.487743
Multiple R-Squared: 0.478546 Adjusted R-Squared: 0.443239
这里有人知道如何解决吗？ 提前谢谢！