R-lrm(rms软件包)错误-无法使用lrm.fit拟合模型

时间:2019-10-18 12:58:34

标签: r

我能够运行以下模型:

> mod1b <- lrm(dependent_variable ~ quarterYear.x, data = merged_dataset)
> print(mod1b)
Logistic Regression Model

 lrm(formula = dependent_variable ~ quarterYear.x, data = merged_dataset)

                       Model Likelihood     Discrimination    Rank Discrim.    
                          Ratio Test           Indexes           Indexes       
 Obs         28198    LR chi2    3466.57    R2       0.181    C       0.714    
  0          22333    d.f.            38    g        0.792    Dxy     0.428    
  1           5865    Pr(> chi2) <0.0001    gr       2.207    gamma   0.438    
 max |deriv| 5e-13                          gp       0.141    tau-a   0.141    
                                            Brier    0.141                     

                      Coef    S.E.   Wald Z Pr(>|Z|)
 Intercept            -1.6422 0.1410 -11.64 <0.0001 
 quarterYear.x=2008 3 -0.0846 0.1980  -0.43 0.6691  
 quarterYear.x=2008 4  0.0553 0.1864   0.30 0.7669  
 quarterYear.x=2009 1  0.1120 0.2210   0.51 0.6122  
 quarterYear.x=2009 2  0.2401 0.2154   1.11 0.2650  
 quarterYear.x=2009 3  0.3360 0.2106   1.60 0.1106  
 quarterYear.x=2009 4  0.4365 0.2076   2.10 0.0355  
 quarterYear.x=2010 1  1.9436 0.1692  11.49 <0.0001 
 quarterYear.x=2010 2  2.0445 0.1673  12.22 <0.0001 
 quarterYear.x=2010 3  1.8433 0.1688  10.92 <0.0001 
 quarterYear.x=2010 4  1.9154 0.1620  11.82 <0.0001 
 quarterYear.x=2011 1  2.0954 0.1741  12.04 <0.0001 
 quarterYear.x=2011 2  1.8132 0.1664  10.89 <0.0001 
 quarterYear.x=2011 3  1.6494 0.1644  10.03 <0.0001 
 quarterYear.x=2011 4  1.7140 0.1669  10.27 <0.0001 
 quarterYear.x=2012 1  0.9454 0.1756   5.38 <0.0001 
 quarterYear.x=2012 2  0.7768 0.1749   4.44 <0.0001 
 quarterYear.x=2012 3  0.9222 0.1732   5.32 <0.0001 
 quarterYear.x=2012 4  0.7500 0.1784   4.20 <0.0001 
 quarterYear.x=2013 1 -0.1136 0.1835  -0.62 0.5360  
 quarterYear.x=2013 2 -0.0625 0.1817  -0.34 0.7308  
 quarterYear.x=2013 3 -0.0304 0.1811  -0.17 0.8667  
 quarterYear.x=2013 4  0.5335 0.1602   3.33 0.0009  
 quarterYear.x=2014 1 -0.0936 0.1680  -0.56 0.5774  
 quarterYear.x=2014 2 -0.0243 0.1679  -0.14 0.8850  
 quarterYear.x=2014 3 -0.0980 0.1689  -0.58 0.5619  
 quarterYear.x=2014 4 -0.1459 0.1684  -0.87 0.3862  
 quarterYear.x=2015 1 -0.2017 0.1678  -1.20 0.2294  
 quarterYear.x=2015 2 -0.4631 0.1712  -2.70 0.0068  
 quarterYear.x=2015 3 -0.3057 0.1675  -1.83 0.0680  
 quarterYear.x=2015 4 -0.3076 0.1666  -1.85 0.0649  
 quarterYear.x=2016 1 -0.1623 0.1638  -0.99 0.3216  
 quarterYear.x=2016 2 -0.3689 0.1670  -2.21 0.0271  
 quarterYear.x=2016 3 -0.7155 0.1744  -4.10 <0.0001 
 quarterYear.x=2016 4 -0.4090 0.1681  -2.43 0.0150  
 quarterYear.x=2017 1 -0.3493 0.1657  -2.11 0.0351  
 quarterYear.x=2017 2 -0.5273 0.1704  -3.10 0.0020  
 quarterYear.x=2017 3 -0.4696 0.1683  -2.79 0.0053  
 quarterYear.x=2017 4 -0.3828 0.1735  -2.21 0.0273

> mod1b <- lrm(dependent_variable ~ var1 + var2 + var3 + var4 +  var5 + var6 + var7 + var8, data = merged_dataset)
> print(mod1b)
Logistic Regression Model

 lrm(formula = dependent_variable ~ var1 + var2 + var3 + var4 +  var5 + var6 + var7 + var8, data = merged_dataset)

                       Model Likelihood     Discrimination    Rank Discrim.    
                          Ratio Test           Indexes           Indexes       
 Obs         28198    LR chi2    2431.49    R2       0.129    C       0.706    
  0          22333    d.f.            11    g        0.833    Dxy     0.412    
  1           5865    Pr(> chi2) <0.0001    gr       2.300    gamma   0.412    
 max |deriv| 0.001                          gp       0.132    tau-a   0.136    
                                            Brier    0.150                     

                  Coef    S.E.    Wald Z Pr(>|Z|)
 Intercept        -1.4032  0.2504  -5.60 <0.0001 
 var1              3.3649  0.0866  38.84 <0.0001 
 var2=EUR         0.0620  0.2449   0.25 0.8000  
 var2=GBP         -5.1954 28.7069  -0.18 0.8564  
 var2=JPY         -0.3392  0.4959  -0.68 0.4939  
 var2=USD         -0.6195  0.2634  -2.35 0.0187  
 var3             -0.2103  0.0387  -5.44 <0.0001 
 var4             0.1288  0.0275   4.68 <0.0001 
 var5             -0.0812  0.0202  -4.02 <0.0001 
 var6             -6.7443  0.3803 -17.73 <0.0001 
 var7             1.5258  0.2817   5.42 <0.0001 
 var8             -0.1011  0.0173  -5.84 <0.0001 

但是下面的一个是前两个的组合给我一个错误:

> mod1b <- lrm(dependent_variable ~ var1 + var2 + var3 + var4 +  var5 + var6 + var7 + var8 + quarterYear.x, data = merged_dataset, na.action="na.delete")
singular information matrix in lrm.fit (rank= 49 ).  Offending variable(s):
quarterYear.x=2017 4 
Warning message:
In lrm(dependent_variable ~ var1 + var2 + var3 +  :
  Unable to fit model using “lrm.fit”
> print(mod1b)
Error in 1:ns : argument of length 0

您知道为什么会发生这种情况以及如何解决此问题吗?

0 个答案:

没有答案