运行逐步回归时出错

时间:2014-11-14 18:26:18

标签: r regression logistic-regression glm

我正在尝试运行逐步回归模型。我一直收到这条消息:

#Error in step(cdc.fit, direction = "backward") : 
#  number of rows in use has changed: remove missing values?
#In addition: There were 50 or more warnings (use warnings() to see the first 50)

我是否因为缺少值而收到此消息?

这是我的代码:

model=glm(health~
      ALCDAY5+
      AVEDRNK2+
      CHILDREN+
      CHKHEMO3+
      POORHLTH+
      BLOODCHOYes+
      BPHIGH4No+
      CHCOCNCRYes
      , data=data, fmaily=binomial)

 stepmodel_back <- model(cdc.fit,direction='backward') 
 summary(stepmodel_back) 

谢谢!

1 个答案:

答案 0 :(得分:0)

我没有在MASS中使用stepAIC()函数,但看起来以下行不正确:

stepmodel_back <- step(cdc.fit, direction = "backward")

您已将glm模型分配给名为&#39; model&#39;的对象,但此行不会引用此对象。应该是:

stepmodel_back <- step(model, direction = "backward)

希望这有帮助。